Cardiovascular Outcomes According to Urinary Albumin and Kidney Disease in Patients With Type 2 Diabetes at High Cardiovascular Risk

Key Points

Question

What is the incremental prognostic value of urinary albumin excretion for cardiovascular risk assessment in patients with diabetes with and without the incorporation of cardiac biomarkers?

Findings

In this secondary analysis of a randomized clinical trial population of 15 760 patients with type 2 diabetes and high cardiovascular risk, there was a stepwise increased risk of cardiovascular events according to baseline urinary albumin to creatinine ratio. The association with increased risk was present even at low-level elevations of urinary albumin to creatinine ratio, which are otherwise considered normal, although this relationship was attenuated when adjusted for cardiac biomarkers.

Meaning

Low levels of albuminuria improve risk stratification for future cardiovascular risk in patients with type 2 diabetes; urinary albumin to creatinine ratio provides minimal incremental prognostic utility beyond cardiac biomarkers.

Abstract

Importance

An elevated level of urinary albumin to creatinine ratio (UACR) is a marker of renal dysfunction and predictor of kidney failure/death in patients with type 2 diabetes. The prognostic use of UACR in established cardiac biomarkers is not well described.

Objective

To evaluate whether UACR offers incremental prognostic benefit beyond risk factors and established plasma cardiovascular biomarkers.

Design, Setting, and Participants

The Saxagliptin Assessment of Vascular Outcomes Recorded in Patients With Diabetes Mellitus–Thrombolysis in Myocardial Infarction (SAVOR-TIMI) 53 study was performed from May 2010 to May 2013 and evaluated the safety of saxagliptin vs placebo in patients with type 2 diabetes with overt cardiovascular disease or multiple risk factors. Median follow-up was 2.1 years (interquartile range, 1.8-2.3 years).

Interventions

Patients were randomized to saxagliptin vs placebo plus standard care.

Main Outcomes and Measures

Baseline UACR was measured in 15 760 patients (95.6% of the trial population) and categorized into thresholds.

Results

Of 15 760 patients, 5205 were female (33.0%). The distribution of UARC categories were: 5805 patients (36.8%) less than 10 mg/g, 3891 patients (24.7%) at 10 to 30 mg/g, 4426 patients (28.1%) at 30 to 300 mg/g, and 1638 patients (10.4%) at more than 300 mg/g. When evaluated without cardiac biomarkers, there was a stepwise increase with each higher UACR category in the incidence of the primary composite end point (cardiovascular death, myocardial infarction, or ischemic stroke) (3.9%, 6.9%, 9.2%, and 14.3%); cardiovascular death (1.4%, 2.6%, 4.1%, and 6.9%); and hospitalization for heart failure (1.5%, 2.5%, 4.0%, and 8.3%) (adjusted P < .001 for trend). The net reclassification improvement at the event rate for each end point was 0.081 (95% CI, 0.025 to 0.161), 0.129 (95% CI, 0.029 to 0.202), and 0.056 (95% CI, −0.005 to 0.141), respectively. The stepwise increased cardiovascular risk associated with a UACR of more than 10 mg/g was also present within each chronic kidney disease category. The UACR was associated with outcomes after including cardiac biomarkers. However, the improvement in discrimination and reclassification was attenuated; net reclassification improvement at the event rate was 0.022 (95% CI, −0.022 to 0.067), −0.008 (−0.034 to 0.053), and 0.043 (−0.030 to 0.052) for the primary end point, cardiovascular death, and hospitalization for heart failure, respectively.

Conclusions and Relevance

In patients with type 2 diabetes, UACR was independently associated with increased risk for a spectrum of adverse cardiovascular outcomes. However, the incremental cardiovascular prognostic value of UACR was minimal when evaluated together with contemporary cardiac biomarkers.

Trial Registration

clinicaltrials.gov Identifier: NCT01107886

This secondary analysis of a randomized clinical trial evaluates the association between urinary albumin to creatine ratios and mortality rates as well as cardiovascular events.

Introduction

Chronic kidney disease is a well-recognized complication of type 2 diabetes (T2D) affecting 30% to 40% of patients.¹ Chronic kidney disease may be recognized by 2 distinct and complementary methods: estimated glomerular filtration rates (eGFR) and urinary albumin to creatinine ratio (UACR).^2,3,4 Elevated levels of urinary albumin, as assessed by UACR, reflect damage to the basement membrane and endothelium of glomerular capillaries and denote the presence of chronic kidney disease, even within different eGFR categories.^2,3 In patients with T2D, UACR often represents diabetic nephropathy, although other diseases, such as hypertension may also contribute. Urinary albumin to creatinine ratio levels between 30 mg/g to 300 mg/g, formerly termed microalbuminuria, represent moderately increased levels of albuminuria (Kidney Disease: Improving Global Outcomes category A2).^2,3,5 In healthy adults, UACR is typically less than 10 mg/g; however, even small elevations in urinary albumin between 10 mg/g and 29 mg/g have been associated with progression of renal disease and increased mortality.^6,7,8,9,10 The incremental value of UACR for the prediction of cardiovascular risk when combined with established cardiovascular biomarkers, such as high-sensitivity troponin T (hsTnT), natriuretic peptides, or high-sensitivity C-reactive protein (hsCRP), has not been well described, to our knowledge.

The Saxagliptin Assessment of Vascular Outcomes Recorded in Patients With Diabetes Mellitus–Thrombolysis in Myocardial Infarction (SAVOR-TIMI) 53 trial¹¹ evaluated the cardiovascular efficacy and safety of saxagliptin, a selective dipeptidyl peptidase-4 inhibitor, in 16 492 patients with T2D with overt atherosclerotic vascular disease or at risk for cardiovascular events. During a median follow-up of 2.1 years, saxagliptin did not alter the risk of the primary composite end point of cardiovascular death, myocardial infarction, or ischemic stroke, although there was a 27% increased relative risk of hospitalization for heart failure in patients assigned to receive saxagliptin.¹² In addition, saxagliptin improved UACR over time compared with placebo.^13,14As part of a prespecified analysis, we evaluated the cardiovascular risk associated with baseline UACR and eGFR together with cardiac biomarkers.

Methods

Study Design and Oversight

The SAVOR-TIMI 53 study¹¹ was a multicenter, randomized double-blind, placebo-controlled trial that randomized patients with T2D, hemoglobin A_1C level between 6.5% and 12.0% within 6 months of randomization, and either a history of established atherosclerotic vascular disease or multiple risk factors for vascular disease (ie, investigator-reported dyslipidemia, hypertension, or smoking) to receive either 5 mg of saxagliptin daily (or 2.5 mg daily in patients with an eGFR of ≤50 mL/min/1.73 m²) or matching placebo.¹¹ The protocol specified the inclusion of at least 800 patients with at least moderate to severe kidney impairment (eGFR, <50 mL/min/1.73 m²), of whom 300 patients were to have an eGFR less than 30 mL/min/1.73 m². Patients with a history of either end-stage renal disease receiving chronic dialysis, serum creatinine level of more than 6.0 mg/dL, or previous kidney transplant were excluded. The full eligibility criteria and analysis plan have been reported previously.^11,13 The trial protocol was reviewed and approved by all relevant ethics committees. Written informed consent was obtained from all patients.

End Points

The primary end point of the trial was a composite of the first occurrence of cardiovascular death, myocardial infarction, or ischemic stroke. The secondary composite end point included the elements of the primary end point and hospitalizations for heart failure, unstable angina, or coronary revascularization. A clinical events committee, unaware of the study group assignments, adjudicated all components of the primary and secondary composite efficacy end points^11,13 using definitions based on draft guidelines for the standardization of end points in cardiovascular trials proposed by the US Food and Drug Administration.¹⁵

Baseline Kidney Function Assessment

The eGFR was determined according to the Modification of Diet in Renal Disease formula based on serum creatinine and categorized as more than 60, 30 to 60, and less than 30 mL/min/1.73 m². Urinary albumin to creatinine ratio was measured from a single voided urine sample by a central laboratory (albumin, lower detection limit of 3 mg/L; creatinine, Jaffe reaction, lower detection limit of 4.0 mg/dL). The lowest reportable level of UACR was 1.0 mg/g. Urinary albumin to creatinine ratio was prospectively categorized as less than 10mg/g, 10 mg/g to 29 mg/g, 30 mg/g to 300 mg/g, and more than 300 mg/g.

Statistical Analysis

Categorical variables were compared using χ² test and continuous variables with a Kruskal-Wallis test. Event rates are presented as 2-year Kaplan-Meier estimates. Estimated glomerular filtration rates and UACR were analyzed as continuous variables (as restricted cubic splines) and then based on the prespecified end points noted above. Multivariable models evaluating the association between UACR and clinical outcomes were adjusted for the following baseline variables: treatment arms (saxagliptin vs placebo), age (continuous), sex, race/ethnicity (white vs nonwhite), history of heart failure, duration of T2D (<5 years, 5-9 years, 10-14 years, 15-19 years, and ≥20 years), hemoglobin A_1C level (continuous), systolic blood pressure (continuous), prior myocardial infarction, history of hypertension, history of dyslipidemia, current smoker, and eGFR (continuous). The multivariable analyses were repeated with the inclusion of baseline N-terminal pro B-type natriuretic peptide (NT-proBNP), hsTnT, and hsCRP in the clinical model in analyses restricted to the 12 177 patients with data available on all 3 biomarkers. Biomarkers were examined as log-transformed as well as categorical variables using quartiles (NTproBNP and hsTnT) and categories for hsCRP (<1 mg/l, 1-3 mg/l, >3 mg/l) when comparing categories of UACR. All models were calibrated using slight variations of variables in the Cox proportional hazards model with and without biomarker data. Calibration was evaluated by deciles of predicted probabilities.¹⁶ Estimates of the C statistic for the clinical model created from previously listed variables were calculated based on the Harrell method¹⁷ and then compared with the models after the addition of the different biomarkers. The discriminative value of the biomarkers was further examined with the method described by Pencina and colleagues¹⁸ to determine the net reclassification improvement (NRI) at the event rate along with 95% CIs based on bootstrap resampling and integrated discrimination improvement.¹⁹

Results

Of 16 492 patients in the SAVOR-TIMI 53 trial, baseline UACR was available in 15 760 patients (95.6%). In total, 5205 patients were female (33.0%). The median UACR was 17.0 mg/g (interquartile range, 6.0-68.0 mg/g). Overall, 5805 patients had a UACR less than 10 mg/g (36.8%); 3891 (24.7%), 10 mg/g to 29 mg/g; 4426 (28.1%), 30 mg/g to 300 mg/g; and 1638 (10.4%), more than 300 mg/g. Patients with higher UACR were more likely not to be white, have higher baseline systolic blood pressures, longer duration of diabetes, a higher prevalence of dyslipidemia, hypertension, and established atherosclerotic disease, more elevated hemoglobin A_1C levels, and lower eGFR. In addition, patients with higher UACR had higher concentrations of hsTnT, NT-proBNP, and hsCRP (eTable 1 in the Supplement).

Baseline UACR and Cardiovascular Outcomes

When evaluated without cardiac biomarkers, there was a stepwise increase in the incidence of all cardiovascular events according to baseline UACR categories (Table 1 and Figure 1). When analyzed as a continuous variable in a multivariable model adjusted for baseline characteristics and eGFR, UACR was significantly associated with an increased risk of all-cause mortality, cardiovascular death, myocardial infarction, and hospitalization for heart failure (Table 2 and Figure 2). When examining the associated risk according to the prespecified UACR categories, there was a consistent, stepwise pattern of increased cardiovascular risk with each level of UACR. Notably, the risk associated with UACR began to increase significantly even in patients with UACR concentrations between 10 mg/g and 29 mg/g (Table 2). The addition of UACR to the clinical model (without biomarkers) significantly improved discrimination and reclassification of risk for all end points with the exception of ischemic stroke and hospitalization for coronary revascularization (Table 3). The NRI at the event rate after the addition of UACR was improved for the primary end point (NRI, 0.081; 95% CI, 0.025 to 0.161), cardiovascular death (NRI, 0.129; 95% CI, 0.029 to 0.202), myocardial infarction (NRI, 0.082; 95% CI, −0.018 to 0.154), and hospitalization for heart failure (NRI, 0.056; 95% CI, −0.005 to 0.141).

Table 1. Rates of Cardiovascular Events (2-Year Kaplan-Meier Estimates) by UACR Category.

Characteristic	UACR Category, No. (%)a
	<10 mg/g (n = 5805)	10 mg/g to 29 mg/g (n = 3891)	30 mg/g to 300 mg/g (n = 4426)	>300 mg/g (n = 1638)	P Valuea
Primary composite end point	232 (3.9)	275 (6.9)	404 (9.2)	241 (14.3)	<.001
Secondary composite end point	515 (8.9)	458 (11.5)	648 (14.7)	365 (22.4)	<.001
Cardiovascular death	86 (1.4)	104 (2.6)	183 (4.1)	121 (6.9)	<.001
Any cause mortality	139 (2.3)	166 (4.0)	268 (5.8)	175 (9.7)	<.001
Fatal/nonfatal myocardial infarction	103 (1.8)	125 (3.1)	183 (4.2)	100 (6.2)	<.001
Ischemic stroke	66 (1.1)	71 (1.8)	98 (2.3)	49 (3.0)	<.001
Hospitalization for heart failure	85 (1.5)	97 (2.5)	177 (4.0)	134 (8.3)	<.001
Hospitalization for coronary revascularization	300 (5.2)	197 (5.0)	246 (5.7)	102 (6.5)	.034

Abbreviation: UACR, urinary albumin to creatinine ratio.

^aUsing a 2-sided log-rank test for trend.

Figure 1. Kaplan-Meier (KM) Estimates According to Baseline Urinary Albumin to Creatinine Ratio Levels .

Table 2. Risk of Cardiovascular Events per Levels of UACR.

End Points	Log of UACR as Continuousa,b		UACR as Categorical Adjusted HR (95% CI)b
	Adjusted HR (95% CI)	P Value	<10 mg/g	10 mg/g to 29 mg/g	30 mg/g to 300 mg/g	>300 mg/g	P Value
Without biomarkers (n = 15 688)
Primary composite end pointc	1.39 (1.32-1.47)	<.001	1 [Reference]	1.65 (1.39-1.97)	2.01 (1.70-2.38)	3.11 (2.55-3.80)	<.001
Secondary composite end pointd	1.33 (1.27-1.39)	<.001	1 [Reference]	1.28 (1.13-1.46)	1.56 (1.38-1.76)	2.35 (2.02-2.72)	<.001
Cardiovascular death	1.55 (1.42-1.70)	<.001	1 [Reference]	1.65 (1.24-2.21)	2.40 (1.84-3.14)	4.10 (3.02-5.57)	<.001
Any cause mortalitye	1.51 (1.40-1.62)	<.001	1 [Reference]	1.66 (1.32-2.09)	2.25 (1.82-2.79)	3.65 (2.85-4.68)	<.001
Myocardial infarctione	1.40 (1.28-1.52)	<.001	1 [Reference]	1.73 (1.33-2.24)	2.10 (1.64-2.69)	2.97 (2.19-4.01)	<.001
Ischemic stroke	1.22 (1.09-1.38)	<.001	1 [Reference]	1.43 (1.02-2.01)	1.56 (1.13-2.16)	1.94 (1.29-2.91)	.001
Hospitalization for heart failuref	1.77 (1.62-1.93)	<.001	1 [Reference]	1.65 (1.22-2.21)	2.68 (2.05-3.51)	5.49 (4.05-7.44)	<.001
Hospitalization for coronary revascularization	1.08 (1.00-1.16)	.045	1 [Reference]	0.98 (0.82-1.18)	1.07 (0.90-1.27)	1.21 (0.94-1.54)	.099
With biomarkers (n = 11 685)
Primary composite end point	1.15 (1.07-1.23)	<.001	1 [Reference]	1.36 (1.12-1.66)	1.33 (1.10-1.62)	1.65 (1.31-2.09)	<.001
Secondary composite end point	1.12 (1.06-1.19)	<.001	1 [Reference]	1.10 (0.95-1.27)	1.11 (0.96-1.28)	1.42 (1.19-1.70)	<.001
Cardiovascular death	1.20 (1.08-1.33)	<.001	1 [Reference]	1.12 (0.81-1.54)	1.30 (0.96-1.76)	1.81 (1.27-2.57)	<.001
Any cause mortality	1.24 (1.14-1.35)	<.001	1 [Reference]	1.18 (0.91-1.54)	1.32 (1.03-1.69)	1.91 (1.43-2.55)	<.001
Myocardial infarctione	1.18 (1.07-1.31)	.001	1 [Reference]	1.60 (1.19-2.16)	1.62 (1.21-2.17)	1.88 (1.31-2.68)	<.001
Ischemic stroke	0.99 (0.86-1.15)	.917	1 [Reference]	1.20 (0.82-1.75)	1.15 (0.79-1.67)	0.99 (0.60-1.63)	.931
Hospitalization for heart failure	1.30 (1.17-1.44)	<.001	1 [Reference]	0.94 (0.67-1.32)	1.28 (0.94-1.75)	1.99 (1.40-2.84)	<.001
Hospitalization for coronary revascularizationg	0.98 (0.90-1.07)	.636	1 [Reference]	0.94 (0.77-1.16)	0.89 (0.72-1.09)	0.93 (0.70-1.25)	.582

Abbreviations: HR, hazard ratio; UACR, urinary albumin to creatinine ratio.

^aPer log (SD) is 1 (1.76).

^bHazard ratio is adjusted for treatment arms (saxagliptin vs placebo), age (years), sex, race (white vs nonwhite), history of heart failure, duration of diabetes (<5 years, 5-9 years, 10-14 years, 15-19 years, and ≥20 years), hemoglobin A_1C (percentage), systolic blood pressure, prior myocardial infarction, history of hypertension, history of dyslipidemia, current smoker, and estimated glomerular filtration rates (millimeters per minute), and for the models with biomarkers, high-sensitivity troponin T quartiles, pro B-type natriuretic peptide quartiles, and C-reactive protein (<1, 1-3, >3 mg/L).

^cUsed a shortened model for when UACR is continuous: age (<65 years vs ≥65 years), race (white vs nonwhite), history of heart failure, hemoglobin A_1C, prior myocardial infarction, hypertension, current smoking, estimated glomerular filtration rates (<30, 30-60, >60 mL/min/1.73 m²).

^dAge (<65 vs ≥65 years), systolic blood pressure (<120, 120-<130, 130-<140, ≥140 mm Hg), and estimated glomerular filtration rates (<30, 30-60, >60 mL/min/1.73 m²) are categorical.

^eAge is binary (<65 vs ≥65 years).

^fAge is binary (<65 vs ≥65 years) only when UACR is categorical. Duration of diabetes was dropped from the model when UACR is continuous.

^gPrior heart failure was dropped from the model.

Figure 2. Risk of Primary End Point, Cardiovascular Death, Myocardial Infarction, and Hospitalization for Heart Failure, by Urinary Albumin to Creatinine Ratio (UACR) as Continuous Variable.

The solid dark blue line indicates the adjusted hazard ratio; dotted dark blue line, 95% CI; and the solid light blue line, a reference when the adjusted hazard ratio for UACR is 1.

Table 3. Improvements in Discrimination and Reclassification of Risk With the Addition of UACR to Clinical Model^a.

End Points	C Statistic Clinical Model Without UACR	C Statistic With UACR as Categorical	Change in C Statistic (95% CI)	LR P Value	Absolute IDI Adding UACR as Categorical	Relative IDI Adding UACR as Categorical	NRI at the Event Rate (95% CI)b
Without biomarkers
Primary composite end point	0.666 (0.650 to 0.682)	0.690 (0.675 to 0.705)	−0.024 (−0.033 to −0.015)	<.001	0.0083 (0.0063 to 0.0103)	0.297	0.081 (0.025 to 0.161)
Secondary composite end pointc	0.655 (0.642 to 0.667)	0.668 (0.656 to 0.681)	−0.014 (−0.020 to −0.008)	<.001	0.0094 (0.0074 to 0.0114)	0.220	0.054 (0.014 to 0.096)
Cardiovascular death	0.718 (0.694 to 0.743)	0.751 (0.728 to 0.773)	−0.033 (−0.046 to −0.019)	<.001	0.0068 (0.0042 to 0.0095)	0.241	0.129 (0.029 to 0.202)
Any cause mortalityd	0.688 (0.668 to 0.709)	0.716 (0.697 to 0.736)	−0.028 (−0.039 to −0.017)	<.001	0.0082 (0.0060 to 0.0104)	0.306	0.127 (0.035 to 0.190)
Myocardial infarctiond	0.676 (0.652 to 0.699)	0.695 (0.672 to 0.718)	−0.020 (−0.031 to −0.008)	<.001	0.0037 (0.0024 to 0.0050)	0.244	0.082 (−0.018 to 0.154)
Ischemic stroke	0.664 (0.631 to 0.696)	0.673 (0.642 to 0.705)	−0.010 (−0.024 to 0.004)	.003	0.0009 (0.0002 to 0.0015)	0.129	−0.014 (−0.080 to 0.138)
Hospitalization for heart failured	0.781 (0.759 to 0.803)	0.807 (0.787 to 0.827)	−0.026 (−0.038 to −0.014)	<.001	0.0159 (0.0110 to 0.0209)	0.284	0.056 (−0.005 to 0.141)
Hospitalization for coronary revascularization	0.629 (0.610 to 0.648)	0.630 (0.611 to 0.649)	−0.001 (−0.004 to 0.001)	.440	0.0002 (−0.0001 to 0.0005)	0.015	0.012 (−0.036 to 0.054)
With biomarkers
Primary composite end point	0.740 (0.724 to 0.756)	0.743 (0.727 to 0.759)	−0.003 (−0.006 to −0.0001)	.001	0.0015 (0.0003 to 0.0028)	0.024	0.022 (−0.022 to 0.067)
Secondary composite end point	0.707 (0.694 to 0.720)	0.708 (0.695 to 0.721)	−0.001 (−0.003 to 0.001)	.006	0.0017 (0.0006 to 0.0028)	0.020	−0.009 (−0.021 to 0.031)
Cardiovascular death	0.815 (0.794 to 0.836)	0.817 (0.797 to 0.838)	−0.003 (−0.006 to 0.0001)	.030	0.0014 (−0.0005 to 0.0033)	0.023	−0.008 (−0.034 to 0.053)
Any cause mortality	0.787 (0.768 to 0.806)	0.790 (0.771 to 0.809)	−0.004 (−0.007 to −0.001)	.003	0.0020 (0.0001 to 0.0039)	0.030	0.015 (−0.034 to 0.049)
Myocardial infarctiond	0.735 (0.711 to 0.759)	0.741 (0.717 to 0.764)	−0.006 (−0.011 to 0.0002)	.005	0.0014 (0.0003 to 0.0024)	0.045	0.051(−0.035 to 0.103)
Ischemic stroke	0.708 (0.674 to 0.742)	0.710 (0.677 to 0.743)	−0.002 (−0.008 to 0.004)	.447	0.0002 (−0.0003 to 0.0007)	0.015	−0.026 (−0.069 to 0.105)
Hospitalization for heart failure	0.875 (0.860 to 0.891)	0.879 (0.863 to 0.894)	−0.003 (−0.006 to −0.001)	<.001	0.0053 (0.0015 to 0.0090)	0.051	0.043 (−0.030 to 0.052)
Hospitalization for coronary revascularizatione	0.655 (0.634 to 0.675)	0.655 (0.635 to 0.676)	−0.0002 (−0.002 to 0.002)	.669	0.0002 (−0.0001 to 0.0005)	0.012	−0.016 (−0.044 to 0.052)

Abbreviations: IDI, integrated discrimination improvement; LR, likelihood ratio; NRI, net reclassification improvement; UACR, urinary albumin to creatinine ratio.

^aThe clinical model includes treatment arms (saxagliptin vs placebo), age (years), sex, race (white vs nonwhite), history of heart failure, duration of diabetes (<5 years, 5-9 years, 10-14 years, 15-19 years, and ≥20 years), hemoglobin A_1C (percentage), systolic blood pressure, prior myocardial infarction, history of hypertension, history of dyslipidemia, current smoker, estimated glomerular filtration rates (milliliters per minute), high-sensitivity troponin T quartiles, pro B-type natriuretic peptide quartiles, and C-reactive protein (<1, 1-3, >3 mg/L) for the models with biomarkers.

^b95% CIs based on bootstrap resampling.

^cAge (<65 vs ≥65 years), systolic blood pressure (<120g, 120-129, 130-139, ≥140 mm Hg) and estimated glomerular filtration rates (<30, 30-60, >60 mL/min/1.73 m²) are categorical.

^dAge is binary (<65 vs ≥65 years).

^ePrior heart failure was dropped from the model.

After adjusting for baseline levels of NT-proBNP, hsTnT, and hsCRP, levels of UACR remained significantly associated with cardiovascular outcomes; however, the relationship was attenuated (Table 2). With the inclusion of cardiac biomarkers, the improvements in C statistics, the NRI at the event rate, and integrated discrimination improvement, while present, were relatively small (Table 3).

Baseline UACR and eGFR

The incidence of cardiovascular events increased with higher concentrations of UACR and lower eGFR, regardless of whether UACR was analyzed as a categorical or continuous variable. Even within each of the 3 different categories of eGFR, a higher UACR was associated with an increased risk of cardiovascular events and therefore identified differential risk within each eGFR category. For example, in patients with normal kidney function or mild kidney insufficiency (eGFR >60 mL/min/1.73 m²), the incidence of the primary end point increased progressively from 3.5% with a UACR of less than 10 mg/g to 12.3% in patients with a UACR of more than 300 mg/g (eTable 2 in the Supplement). This excess risk remained significant after adjusting for other baseline characteristics (eFigure and eTable 3 in the Supplement) and for cardiac biomarkers (eTable 4 in the Supplement).

The association between UACR and outcomes was consistent in patients treated with saxagliptin or placebo (P values for interactions were all >.05, except for ischemic stroke [P = .033]) (eTable 5 in the Supplement) and irrespective of baseline use of angiotensin-converting enzyme inhibitors or angiotensin receptor blockers (eTable 6 in the Supplement).

Discussion

In this study of 15 760 patients with T2D at high cardiovascular risk, we found that when evaluated together with standard clinical markers including eGFR, baseline UACR was independently associated with total mortality as well as cardiovascular events, such as cardiovascular death, myocardial infarction, ischemic stroke, and hospitalization for heart failure, thus expanding prior observations and supporting the hypothesis that UACR provides complementary insight into the association between diabetic kidney disease and cardiovascular risk. Currently, urinary albumin testing is already recommended for all patients with T2D to assess for chronic kidney disease. Urinary albumin to creatinine ratio could therefore readily be used as a more formal tool for cardiovascular risk prognostication without any additional testing or costs above standard therapy.

In contrast, when evaluated simultaneously with 3 frequently used cardiac plasma biomarkers (hsTnT, NT-proBNP, and hsCRP), the incremental prognostic value of UACR was minimal. This is not unexpected given the strong association between cardiac biomarkers and cardiovascular outcomes in patients with T2D.^20,21 To our knowledge, no practice guidelines, however, currently recommend their use in stable patients with diabetes. Consequently, hsTnT, NT-proBNP, and hsCRP are rarely used in this clinical setting, and UACR may therefore still offer additional incremental prognostic information. Eventual integration of cardiac biomarkers into risk stratification algorithms in T2D would provide additional clinical value, although at increased cost, through a more nuanced risk assessment.

In patients with T2D, increased levels of urinary albumin are likely an early signal of microvascular disease and indicate some degree of kidney damage.⁴ Historically, a cut point of 30 mg/g has been used in the diagnosis of albuminuria and thereby labeling a patient with kidney disease, which is diabetic nephropathy in most patients with T2D.^22,2 Our data indicate that from a prognostic stand point, even low-level elevations in UACR (10-30 mg/L), which would not be classified by contemporary clinical standards as elevated levels of albuminuria,⁵ are associated with increased all-cause mortality as well as cardiovascular risk when compared with patients with UACR of less than 10 mg/g and are therefore clinically relevant and successfully identify high-risk patients.

The association between UACR and heart failure has been reported previously in a general population of patients with prevalent heart failure^6,23,24,25 and in patients with T2D.^6,26,27 Without the inclusion of cardiac biomarkers, the risk of heart failure hospitalization in SAVOR-TIMI 53 began to rise with a UACR of more than 10 mg/g, even after adjusting for many baseline characteristics and regardless of baseline eGFR. We have previously demonstrated that in the SAVOR-TIMI 53 population, baseline eGFR is associated with an increased cardiovascular risk, including heart failure²⁸; however, that analysis did not include UACR. Levels of NT-proBNP and hsTnT are strongly associated with worsening heart failure in stable populations, so the attenuation of the association between UACR and hospitalization for heart failure when these cardiac biomarkers were added to the model is not surprising.

The association between albuminuria and cardiovascular mortality and end-stage kidney disease are well described in patients with T2D and in the general population.^{6,7,9,29,30,31} Other studies, many including patients without T2D, found an association between UACR and heart failure,^25,32 coronary heart disease³² and incident hypertension. None of these studies included cardiac biomarkers, which have been described as some of the most robust predictors of risk in primary and secondary prevention populations of patients with T2D.^20,21 Other studies in patients with and without diabetes that also found an association between UACR and cardiovascular events lacked a sufficient number of patients to evaluate the different cardiac events individually³³ or did not simultaneously assess both natriuretic peptides and high-sensitivity troponin.^34,35,36

In addition to inhibitors of the renin-angiotensin-aldosterone system, several glucose-lowering drugs, including dipeptidyl peptidase 4 inhibitors,³⁷ sodium-glucose cotransporter 2 inhibitors,³⁸ and glucagon-like peptide 1 agonist,³⁹ improve UACR. In SAVOR-TIMI 53, saxagliptin improved UACR compared with placebo, regardless of baseline eGFR and UACR.^13,14 The apparent discordance between the minor reductions in UACR with saxagliptin without any corresponding benefit in major adverse cardiovascular events with saxagliptin may be because of a median follow-up of 2 years that may have been sufficient to improve UACR but not sufficient to observe any cardiovascular benefit. Moreover, it is not known whether UACR is a causal vs bystander marker of cardiovascular risk such that lower UACR per se would result in improved outcomes.

Limitations

Urinary albumin to creatinine ratio was only measured once at baseline in this study; thus, we cannot exclude intrapatient sampling variability. However, this imprecision would likely bias toward a weaker association between UACR and outcomes. We did not measure cystatin C in this population and therefore cannot correlate outcomes with this kidney biomarker. Changes after baseline in medications and subsequent changes in glycemic indices or blood pressure were not evaluated in this analysis and therefore cannot account for any differences that might influence the association between UACR and outcomes.

Conclusions

Elevated levels of UACR, even within what has been considered to be in the normal range, are independently associated with increased risk of all-cause mortality as well as across a spectrum of cardiovascular end points, even after adjusting for known cardiovascular risk factors and eGFR. In contrast, the prognostic value of UACR when evaluated in the context of cardiac biomarkers such as natriuretic peptides and high-sensitivity troponin was minimal. Thus the utility of using UACR as a tool for cardiovascular risk assessment in patients with T2D is dependent on whether established cardiac biomarkers are also being assessed simultaneously.

Supplement.

eTable 1. Baseline characteristics

eTable 2. Rates of cardiovascular events (2yr KM%) by eGFR and UACR categories

eTable 3. Risk of cardiovascular events according to UACR and eGFR

eTable 4. Risk of cardiovascular events according to UACR and eGFR (adjusted for baseline biomarkers)

eTable 5. Risk of cardiovascular events by UACR continuous by treatment arms (excluding biomarkers)

eTable 6. Risk of cardiovascular events by UACR continuous by baseline ACE inhibitor/ARB use (excluding biomarkers)

eFigure. Adjusted risk of different endpoints according to UACR and eGFR categories for the primary composite endpoint