Heart and Kidney Outcomes With Ertugliflozin in People with Non-albuminuric Diabetic Kidney Disease: A post hoc Analysis from the Randomized VERTIS CV Trial.

Introduction: Using data from the VERTIS CV trial (NCT01986881), the impact of ertugliflozin in patients with nonalbuminuric diabetic kidney disease (DKD-non-Alb) was assessed.
Methods: Patients with type 2 diabetes mellitus (T2DM) and atherosclerotic cardiovascular disease (ASCVD) were randomized to ertugliflozin or placebo. Subgroups were defined by estimated glomerular filtration rate (eGFR) (ml/min per 1.73 m2) and urinary albumin-to-creatinine ratios (UACRs) (mg/g): DKD-Non-Alb (eGFR < 60 + UACR < 30, n = 867); Alb DKD stage 3 (DKD stage 3 Alb, eGFR < 60 + UACR ≥ 30, n = 891); Alb DKD stages 1 + 2 (DKD stages 1-2 Alb, eGFR ≥ 60 + UACR ≥ 30, n = 2356); and no DKD (non-DKD, eGFR ≥ 60 + UACR < 30, n = 3916). eGFR slopes, eGFR and UACR over time, time to first event of a prespecified exploratory kidney composite outcome, albuminuria progression, and hospitalization for heart failure (HHF) were assessed.
Results: Total eGFR slopes (ml/min per 1.73 m2 per year; weeks 0-260) with placebo were -0.23, -1.27, -2.29, and -1.19 for the DKD-Non-Alb, DKD stage 3 Alb, DKD stages 1 to 2 Alb, and non-DKD subgroups, respectively (P < 0.0001). Similar trends were found with ertugliflozin but with reduced rates of decline. Ertugliflozin treatment resulted in a significant reduction in the risk for albuminuria progression across subgroups, with Alb subgroups having the largest relative risk reduction (P interaction = 0.04). The hazard ratios (HRs) for ertugliflozin revealing reduction in the risk of the exploratory kidney composite outcome versus placebo was consistent across subgroups (P interaction = 0.34). Alb and the DKD-non-Alb subgroups had a larger relative risk reduction in the HHF outcome compared with other subgroups (P interaction = 0.046).
Conclusion: Among the subgroups, participants with DKD-non-Alb had the slowest rate of eGFR decline. Ertugliflozin treatment resulted in reductions in albuminuria and slower decline in eGFR across subgroups. The effect of ertugliflozin on the HHF outcome was larger in those with more advanced kidney disease.

DKD is a leading cause of end-stage kidney disease globally. The classical paradigm for the clinical course of DKD is a progressive natural history characterized by gradually increasing albuminuria and a decline in GFR leading to end-stage kidney disease. Risk markers for DKD, including albuminuria and eGFR impairment, are also associated with the development of ASCVD, HHF, and mortality., There has been growing recognition that some patients with DKD do not follow the classical pathway to end-stage kidney disease characterized by progressive albuminuria and instead exhibit progressive eGFR decline in the presence of normal levels of UACR. Although the mechanisms responsible for this DKD-non-Alb phenotype remain incompletely understood, several factors may be responsible, including the use of medications that block the renin–angiotensin–aldosterone system and thereby attenuate albuminuria. In addition, the DKD-non-Alb clinical phenotype is associated with the development of kidney disease in elderly patients with diabetes, in whom ischemia may contribute to DKD progression. Ischemia may impair kidney function while reducing intraglomerular pressure and associated albuminuria.,,

Regardless of the pathophysiological basis for the DKD-non-Alb clinical presentation, it is important to understand the risk factors linked with DKD-non-Alb to identify those at risk for this condition, and thereby potentially institute earlier therapies. Insights into the natural history of DKD-non-Alb are needed to understand the role for emerging therapies, including SGLT2 inhibitors, which may prevent eGFR loss even in the absence of albuminuria.

To better understand the role of SGLT2 inhibition in patients with DKD-non-Alb, we evaluated the VERTIS CV trial cohort, which included a large proportion of patients with incidental DKD. VERTIS CV was a cardiovascular outcomes trial (NCT01986881) that assessed the effects of ertugliflozin compared with placebo in patients with T2DM and established ASCVD.6, 7, 8 A prespecified exploratory kidney composite outcome (sustained ≥40% decrease in eGFR from baseline, dialysis/transplantation, or kidney death) revealed a 34% relative risk reduction with ertugliflozin versus placebo. The prespecified secondary outcome of time to first HHF event revealed a 30% relative risk reduction with ertugliflozin versus placebo and a 30% reduction in total HHF events., In the VERTIS CV trial, 21.9% of the subjects at baseline had an eGFR <60 ml/min per 1.73 m2. The goal of this post hoc analysis was to evaluate participants with DKD-non-Alb compared with the other subgroups from the VERTIS CV trial to better understand the impact of DKD-non-Alb on subsequent kidney disease progression and HHF and to compare the effects of ertugliflozin across the 4 groups.

Methods

The VERTIS CV trial (protocol MK-8835-004-01; NCT01986881) was a multicenter, randomized, double-blind, placebo-controlled, event-driven trial in patients with T2DM and established ASCVD comparing placebo with 2 doses of ertugliflozin (5 mg and 15 mg). The VERTIS CV trial design, eligibility, primary results, kidney end points, and full study protocol have been previously published.6, 7, 8,

Study Population

The trial recruited patients with T2DM and established ASCVD who had a baseline screening eGFR ≥30 ml/min per 1.73 m2. The trial was approved by the appropriate institutional review boards and regulatory agencies, with all participants providing written informed consent.6, 7, 8,

Classification by Baseline eGFR and Alb Status

In this post hoc analysis, patients required both a baseline serum creatinine measure (needed to estimate GFR using the Modification of Diet in Renal Disease formula) and a baseline UACR measure for subgroup stratification. On the basis of these 2 criteria, patients were assigned to 1 of the following 4 subgroups based on baseline clinical presentation of DKD:

DKD-non-Alb: baseline eGFR <60 ml/min per 1.73 m2 and UACR <30 mg/g;

Alb DKD stage 3 (DKD stage 3 Alb): baseline eGFR <60 ml/min per 1.73 m2 and UACR ≥30 mg/g;

Alb DKD stages 1 and 2 (DKD stages 1–2 Alb): baseline eGFR ≥60 ml/min per 1.73 m2 and UACR ≥30 mg/g; and

no DKD (non-DKD): baseline eGFR ≥60 ml/min per 1.73 m2 and UACR <30 mg/g.

Outcomes

We evaluated the following in all subgroups by treatment: baseline characteristics, eGFR over time (at baseline and weeks 6, 18, 52, 104, 156, 208, and 260), eGFR slopes, and UACR.

The slopes for changes in eGFR per week or per year were analyzed by random coefficient models. Least square mean differences between ertugliflozin (observations from both doses were pooled for all analyses) and placebo for the weekly or yearly eGFR slopes were assessed for the following 5 periods:

acute eGFR “dip” period: weekly slope from week 0 (baseline) to week 6;

post-eGFR “dip” readjustment period: yearly slope from week 6 to week 52;

postadjustment chronic period: yearly slope from week 52 to week 260;

chronic slope: yearly slope from week 6 to week 260; and

total yearly slope from week 0 (baseline) to week 260.

The time to first event of progression of albuminuria (progression from normoalbuminuria [UACR <30 mg/g] to microalbuminuria or macroalbuminuria [UACR 30–300 mg/g or >300 mg/g, respectively], or progression from microalbuminuria to macroalbuminuria) was analyzed for ertugliflozin versus placebo by subgroup. Patients with baseline macroalbuminuria were excluded from the analysis. The time to first event of a kidney composite end point (comprising sustained ≥40% decline in eGFR, chronic kidney replacement therapy [dialysis or transplantation], or kidney death) and the time to first HHF were also analyzed for ertugliflozin versus placebo by subgroup.

Statistical Analysis

The analyses of eGFR and UACR were performed on the full analysis set (randomized participants who received 1 or more doses of blinded study medication and had 1 or more measurements of the analysis end point). The analyses of eGFR and UACR were performed on the pooled ertugliflozin population and placebo. Data after the initiation of glycemic rescue therapy were included; however, data obtained >2 days after the last dose of study medication were excluded from the eGFR and UACR outcomes analyses.

Least squares mean changes from baseline over time by subgroup were estimated using the repeated measures analysis of covariance method. The repeated measures analysis of covariance models adjusted for the baseline value of the outcome variable, baseline glycated hemoglobin (HbA1c), treatment, visit, subgroup, treatment-by-subgroup interaction, and treatment-by-subgroup-by-visit interaction. Visit was treated as a categorical variable. An unstructured covariance matrix was used to model the correlation among repeated measurements.

The eGFR slope models included the eGFR value as a response variable, with treatment, visit, baseline HbA1c, baseline eGFR, and treatment-by-visit interaction as linear covariates. Visit was treated as a continuous variable. The model enabled individual participant slopes to vary by random effects of intercept and time. An unstructured covariance matrix was used to model the correlation of random effects. Missing data were not imputed. The random effects model used a likelihood-based estimation, which produced unbiased estimates for data missing at random. Treatment-by-subgroup interaction was tested by generalized random coefficient models with treatment, time, subgroup, and treatment-by-subgroup interaction as linear covariates. The P values of eGFR slopes by subgroups in the ertugliflozin or placebo groups were estimated by models including eGFR values as response variables, with subgroup, visit, baseline HbA1c, baseline eGFR, and subgroup-by-visit interaction as linear covariates, whereas visit was treated as a continuous variable. The interactions (Pinteraction) of treatment-by-subgroup were tested to determine whether the effect of ertugliflozin versus placebo was modified by the subgroups.

Owing to the non-normal distribution of UACR, UACR data were log transformed before the analysis. The adjusted mean percentage change (derived from exponentiation of adjusted estimates from the repeated measures analysis of covariance model) in UACR with 95% CIs are presented by treatment and time point. The difference between ertugliflozin treatment and placebo in mean percentage change in UACR from baseline was estimated and presented.

The time-to-event outcomes were analyzed using a stratified Cox proportional hazards model, which included treatment, subgroup, and treatment-by-subgroup interaction, with cohort as a stratification factor (cohort 1 [participants randomized before protocol amendment, between December 2013 and July 2015] and cohort 2 [participants randomized after protocol amendment, in 2016 and beyond]). Time-to-event outcome analyses were performed with data from all randomized patients (the intention-to-treat population) for the prespecified exploratory kidney composite end point and time to first HHF. Data from all patients as treated were used for the analysis of time to progression of albuminuria. Patients with baseline UACR >300 mg/g were excluded from the time to progression of albuminuria analysis. No multiplicity adjustment was made for this post hoc analysis. Baseline characteristics were summarized using descriptive statistics. The analyses were performed using SAS software, version 9.4 (SAS Institute, Cary, NC).

Results

Baseline Characteristics

In the VERTIS CV trial, 8030 participants had both baseline eGFR and UACR measurements. At baseline, 867 (10.8%), 891 (11.1%), 2356 (29.3%), and 3916 (48.8%) patients had DKD-non-Alb, DKD stage 3 Alb, DKD stages 1 to 2 Alb, and non-DKD, respectively. Table 1 displays the baseline demographic and clinical characteristics of the 4 evaluated subgroups. The 2 subgroups with baseline eGFR <60 ml/min per 1.73 m2 (DKD-non-Alb and DKD stage 3 Alb) tended to be older, with a longer duration of T2DM, a higher baseline use of insulin and diuretics (including loop diuretics and mineralocorticoid receptor antagonists), and a lower baseline use of biguanides compared with the DKD stages 1 to 2 Alb and non-DKD subgroups. Renin–angiotensin–aldosterone system inhibitor use was similar across the 4 subgroups. The 2 subgroups with elevated albuminuria at baseline (DKD stage 3 Alb and DKD stages 1–2 Alb) tended to have a higher mean HbA1c and systolic blood pressure (SBP) than the non-Alb subgroups (DKD-non-Alb and non-DKD). Alb subgroups had a greater baseline use of insulin than the corresponding non-Alb groups. The DKD-non-Alb subgroup had a higher proportion of female patients, participants who had no history of tobacco use, and patients with coronary artery disease compared with the other 3 subgroups. Compared with the DKD stage 3 Alb subgroup, the DKD-non-Alb subgroup had a lower use of loop diuretics, but a higher use of mineralocorticoid receptor antagonists. Within each subgroup, baseline demographic and clinical characteristics were generally balanced between those randomized to ertugliflozin and placebo (Supplementary Table S1).

Table 1

Baseline demographic and clinical characteristics (intention-to-treat population)

Characteristic	DKD-non-Alb (n = 867) (10.8%)	DKD stage 3 Alb (n = 891) (11.1%)	DKD stages 1–2 Alb (n = 2356) (29.3%)	Non-DKD (n = 3916) (48.8%)
Age, yr	68.7 (7.3)	67.6 (7.8)	63.2 (7.7)	63.4 (7.9)
Age, yr	69.0 (64.0–74.0)	68.0 (63.0–73.0)	63.0 (58.0–68.0)	64.0 (58.0–69.0)
Female, n (%)	363 (41.9)	270 (30.3)	559 (23.7)	1211 (30.9)
BMI, kg/m2	32.6 (5.4)	32.2 (5.6)	32.0 (5.3)	31.6 (5.3)
BMI, kg/m2	31.9 (28.9–35.4)	31.7 (28.4–35.3)	31.5 (28.6–35.2)	31.0 (27.9–34.5)
Duration of T2DM, yr	15.4 (9.4)	16.1 (8.6)	13.0 (8.1)	11.8 (7.8)
Duration of T2DM, yr	14.3 (8.7–20.2)	15.4 (9.7–21.3)	11.6 (7.0–17.2)	10.3 (6.0–15.9)
Current smoker	60 (6.9)	88 (9.9)	383 (16.3)	562 (14.4)
Past smoker	345 (39.8)	393 (44.1)	936 (39.7)	1560 (39.8)
Never smoked	462 (53.3)	410 (46.0)	1036 (44.0)	1794 (45.8)
HbA1c, %	8.1 (0.9)	8.3 (0.9)	8.4 (1.0)	8.1 (0.9)
HbA1c, %	8.0 (7.5–8.7)	8.2 (7.7–8.9)	8.3 (7.6–9.1)	8.0 (7.4–8.8)
UACR, mg/g	8.0 (4.0–16.0)	125.0 (59.0–419.0)	84.0 (47.0–226.5)	8.0 (4.0–14.0)
eGFR, ml/min per 1.73 m2 (MDRD)	49.8 (7.7)	47.9 (8.2)	83.3 (16.9)	83.6 (16.4)
eGFR, ml/min per 1.73 m2 (MDRD)	51.0 (45.0–56.0)	49.0 (42.0–55.0)	81.0 (70.0–94.0)	81.0 (71.0–93.0)
SBP, mm Hg	130.2 (14.5)	136.8 (14.5)	136.3 (13.5)	131.6 (13.1)
SBP, mm Hg	130.8 (121.0–140.3)	136.7 (127.7–147.0)	136.2 (127.3–145.3)	131.7 (123.0–140.7)
Hb, g/dl	13.5 (1.4)	13.5 (1.5)	14.2 (1.4)	14.1 (1.3)
Hb, g/dl	13.5 (12.6–14.5)	13.4 (12.5–14.5)	14.2 (13.3–15.1)	14.1 (13.3–14.9)
Serum triglyceride, mg/dl	178.0 (123.2)	196.3 (114.5)	191.0 (124.3)	171.8 (106.4)
Serum triglyceride, mg/dl	154.0 (114.0–209.0)	168.0 (123.0–241.0)	159.5 (115.0–226.0)	146.0 (106.0–207.0)
Serum LDL-C, mg/dl	85.7 (33.8)	88.9 (41.0)	90.3 (39.3)	89.2 (37.8)
Serum LDL-C, mg/dl	79.0 (61.0–104.0)	81.0 (60.0–109.0)	82.9 (62.0–112.0)	82.0 (61.0–110.0)
Serum HDL-C, mg/dl	43.8 (12.0)	42.9 (12.2)	43.1 (12.3)	44.4 (12.0)
Serum HDL-C, mg/dl	42.5 (36.0–50.0)	40.5 (35.0–50.0)	41.0 (35.0–50.0)	43.0 (36.0–51.0)
Glucose-lowering agents
Metformin	560 (64.6)	535 (60.0)	1895 (80.4)	3148 (80.4)
Insulin	475 (54.8)	558 (62.6)	1157 (49.1)	1603 (40.9)
Antihypertensive agents	846 (97.6)	873 (98.0)	2231 (94.7)	3701 (94.5)
RAAS inhibitors	709 (81.8)	749 (84.1)	1929 (81.9)	3129 (79.9)
Diuretics	489 (56.4)	507 (56.9)	947 (40.2)	1517 (38.7)
Loop diuretics	219 (25.3)	258 (29.0)	317 (13.5)	431 (11.0)
MRA	123 (14.2)	97 (10.9)	155 (6.6)	283 (7.2)
Lipid-modifying agents	756 (87.2)	774 (86.9)	1933 (82.0)	3323 (84.9)
History of CAD	712 (82.1)	673 (75.5)	1723 (73.1)	2978 (76.0)
History of HF	221 (25.5)	244 (27.4)	573 (24.3)	873 (22.3)

BMI, body mass index; CAD, coronary artery disease; DKD-non-Alb, nonalbuminuric diabetic kidney disease; DKD stages 1–2 Alb, albuminuric stages 1 and 2 diabetic kidney disease; DKD stage 3 Alb, albuminuric stage 3 diabetic kidney disease; eGFR, estimated glomerular filtration rate; Hb, hemoglobin; HbA1c, glycated hemoglobin; HDL-C, high-density lipoprotein cholesterol; HF, heart failure; LDL-C, low-density lipoprotein cholesterol; MDRD, Modification of Diet in Renal Disease; MRA, mineralocorticoid receptor antagonist; Non-DKD, nondiabetic kidney disease; RAAS, renin–angiotensin–aldosterone system; SBP, systolic blood pressure; T2DM, type 2 diabetes mellitus; UACR, urinary albumin-to-creatinine ratio.

Participants required both baseline eGFR and UACR values. Values are mean (SD), n (%), or median (interquartile range).

Acute eGFR Dip Period—Weekly Changes in eGFR From Weeks 0 to 6

During the acute eGFR “dip” period (week 0 [baseline] to week 6), placebo-treated subgroups with eGFR <60 ml/min per 1.73 m2 (DKD-non-Alb and DKD stage 3 Alb) had an increase in mean eGFR, whereas placebo-treated subgroups with eGFR ≥60 ml/min per 1.73 m2 (DKD stages 1–2 Alb and non-DKD) had a decrease in mean eGFR. Acute eGFR slopes (ml/min per 1.73 m2 per week [95% CI]) in those randomized to placebo were 0.53 (0.35–0.71), 0.24 (0.07–0.41), −0.14 (−0.32 to 0.03), and −0.23 (−0.37 to −0.08) in the DKD-non-Alb, DKD stage 3 Alb, DKD stages 1 to 2 Alb, and non-DKD subgroups, respectively (P < 0.0001; Figure 1 and Figure 2a and Supplementary Tables S2 and S3).

Figure 1

Mean eGFR over time in (a) the subgroups with baseline eGFR <60 ml/min per 1.73 m2 (b) and in the groups with baseline eGFR ≥60 ml/min per 1.73 m2 (FAS population). Participants required both baseline eGFR and UACR values for inclusion in the analysis. BL, baseline; DKD-non-Alb, nonalbuminuric diabetic kidney disease; DKD stages 1–2 Alb, albuminuric stages 1 and 2 diabetic kidney disease; DKD stage 3 Alb, albuminuric stage 3 diabetic kidney disease; eGFR, estimated glomerular filtration rate; FAS, full analysis set; non-DKD, nondiabetic kidney disease; UACR, urinary albumin-to-creatinine ratio.

Figure 2

Weekly eGFR slope in the acute “dip” period (week 0 to week 6) in (a) patients randomized to placebo, (b) patients randomized to ertugliflozin, with the (c) placebo-adjusted values illustrated (FAS population). Participants required both baseline eGFR and UACR values for inclusion in the analysis. CI, confidence interval; DKD-non-Alb, nonalbuminuric diabetic kidney disease; DKD stages 1–2 Alb, albuminuric stages 1 and 2 diabetic kidney disease; DKD stage 3 Alb, albuminuric stage 3 diabetic kidney disease; eGFR, estimated glomerular filtration rate; FAS, full analysis set; LSM, least squares mean; non-DKD, nondiabetic kidney disease; UACR, urinary albumin-to-creatinine ratio; W, week.

In patients treated with ertugliflozin, those with eGFR <60 ml/min per 1.73 m2 had slower rates of eGFR decline during the acute eGFR “dip” period, which was slowest in the DKD-non-Alb subgroup, compared with the DKD stages 1 to 2 Alb and non-DKD subgroups (P < 0.0001; Figure 1 and Figure 2b and Supplementary Tables S2 and S3). Treatment with ertugliflozin was associated with a consistent reduction in eGFR compared with placebo in all subgroups (with a similar effect across subgroups), in line with what is known about SGLT2 inhibitor-induced acute hemodynamic effects (Pinteraction = 0.68, Figure 1 and Figure 2c and Supplementary Tables S2 and S3).

Post-eGFR Dip Readjustment Period—Yearly Changes in eGFR From Weeks 6 to 52

During the post-eGFR “dip” readjustment period (weeks 6 to 52), eGFR trajectories were generally unchanged in all placebo-treated subgroups, with yearly eGFR slopes close to 0 (P = 0.82; Supplementary Figure S1A). Mean eGFR increased during weeks 6 to 52 in all ertugliflozin-treated subgroups, with a higher increase observed in subgroups with a baseline UACR <30 mg/g (DKD-non-Alb and non-DKD) compared with other subgroups. eGFR slopes (ml/min per 1.73 m2 per year [95% CI]) in those randomized to ertugliflozin were 2.47 (1.44–3.50), 0.64 (−0.42 to 1.70), 0.83 (−0.01 to 1.67), and 2.12 (1.53 to 2.71) in the DKD-non-Alb, DKD stage 3 Alb, DKD stages 1 to 2 Alb, and non-DKD subgroups, respectively (P = 0.01; Supplementary Figure S1B and Supplementary Table S3). Placebo-adjusted eGFR slopes during weeks 6 to 52 reflected the increase in eGFR during weeks 6 to 52 with ertugliflozin, with a similar effect across subgroups (Pinteraction = 0.88; Supplementary Figure S1C).

Postadjustment Chronic Period—Yearly Changes in eGFR From Weeks 52 to 260

During the postadjustment chronic period (weeks 52 to 260), a slower rate of yearly eGFR decline was observed in subgroups with a baseline UACR <30 mg/g (DKD-non-Alb and non-DKD, with the slowest yearly eGFR decline in the DKD-non-Alb subgroup) compared with other subgroups, regardless of treatment allocation (P ≤ 0.0001; Figure 1, Supplementary Figure S2A and B and Supplementary Table S3). Treatment with ertugliflozin was associated with a slower rate of eGFR decline compared with placebo in all subgroups, without interaction by subgroup (Pinteraction = 0.90; Supplementary Figure S2C and Supplementary Table S3).

Total and Chronic eGFR Slopes—Yearly Changes in eGFR From Weeks 0 or 6 to 260

During weeks 0 to 260, compared with the other subgroups, the DKD-non-Alb subgroup had the slowest rate of yearly eGFR decline, whereas the DKD stages 1 to 2 Alb subgroup had the fastest rate of yearly eGFR decline, regardless of treatment allocation (P < 0.0001; Figure 3a and b). In patients treated with placebo, total eGFR slopes (ml/min per 1.73 m2 per year [95% CI]) were −0.23 (−0.51 to 0.04), −1.27 (−1.57 to −0.96), −2.29 (−2.56 to −2.02), and −1.19 (−1.37 to −1.01) in the DKD-non-Alb, DKD stage 3 Alb, DKD stages 1 to 2 Alb, and non-DKD subgroups, respectively (Figure 3a and Supplementary Table S3). Treatment with ertugliflozin resulted in a significantly slower rate of yearly eGFR decline in all subgroups compared with placebo without interaction by subgroup; the placebo-adjusted eGFR slopes (ml/min per 1.73 m2 per year [95% CI]) during weeks 0 to 260 were 0.55 (0.22–0.88), 0.79 (0.41–1.16), 1.23 (0.91–1.55), and 0.96 (0.73–1.18) in the DKD-non-Alb, DKD stage 3 Alb, DKD stages 1 to 2 Alb, and non-DKD subgroups, respectively (Pinteraction = 0.85; Figure 3c and Supplementary Table S3). Overall, similar results were found with yearly chronic eGFR slopes (weeks 6 to 260; Supplementary Figure S3A–C and Supplementary Table S3).

Figure 3

Yearly eGFR slope in the total period (week 0 to week 260) in (a) patients randomized to placebo, (b) patients randomized to ertugliflozin, (c) with the placebo-adjusted values illustrated (FAS population). Participants required both baseline eGFR and UACR values for inclusion in the analysis. CI, confidence interval; DKD-non-Alb, nonalbuminuric diabetic kidney disease; DKD stages 1–2 Alb, albuminuric stages 1 and 2 diabetic kidney disease; DKD stage 3 Alb, albuminuric stage 3 diabetic kidney disease; eGFR, estimated glomerular filtration rate; FAS, full analysis set; LSM, least squares mean; non-DKD, nondiabetic kidney disease; UACR, urinary albumin-to-creatinine ratio; W, week.

Changes in UACR and Progression of Albuminuria

At year 5, percent change from baseline in UACR (95% CI) with ertugliflozin compared with placebo was −17.0% (−39.5 to 13.8), −14.5% (−39.2 to 20.1), −28.5% (−40.2 to −14.6), and −8.0% (−19.5 to 5.0) in the DKD-non-Alb, DKD stage 3 Alb, DKD stages 1 to 2 Alb, and non-DKD subgroups, respectively (Supplementary Figure S4A–D). Ertugliflozin was associated with a reduced risk of progression of albuminuria in all subgroups compared with placebo, with the greatest relative risk reduction in the Alb subgroups (Supplementary Table S4). The HRs (95% CIs) were 0.79 (0.63–0.99), 0.68 (0.47–0.98), 0.61 (0.49–0.76), and 0.86 (0.77–0.96) in the DKD-non-Alb, DKD stage 3 Alb, DKD stages 1 to 2 Alb, and non-DKD subgroups, respectively (Pinteraction = 0.04).

Prespecified Exploratory Kidney Composite Outcome

In the overall population, treatment with ertugliflozin compared with placebo was associated with a 34% relative reduction in the risk for the prespecified exploratory kidney composite outcome of sustained ≥40% reduction from baseline in eGFR, chronic kidney dialysis (CKD)/transplant, or kidney death, with a HR (95% CI) of 0.66 (0.50–0.88). The impact of ertugliflozin on the prespecified exploratory kidney composite outcome did not differ across subgroups compared with placebo. The HRs (95% CIs) were 0.82 (0.20–3.44), 0.90 (0.49–1.68), 0.66 (0.43–1.02), and 0.44 (0.26–0.74) in the DKD-non-Alb, DKD stage 3 Alb, DKD stages 1 to 2 Alb, and non-DKD subgroups, respectively (Pinteraction = 0.34; Supplementary Table S5). In this analysis, the DKD-non-Alb subgroup had the lowest event rate for the prespecified exploratory kidney composite outcome compared with the other subgroups.

HHF Outcome

In the overall population, the prespecified secondary outcome of time to first HHF had a 30% relative risk reduction with ertugliflozin versus placebo, with a HR (95% CI) of 0.70 (0.54–0.90). Ertugliflozin was associated with a reduced relative risk of time to first HHF in higher risk subgroups (DKD-non-Alb, DKD stage 3 Alb, and DKD stages 1–2 Alb) compared with placebo. The HRs (95% CIs) were 0.72 (0.31–1.69), 0.45 (0.28–0.73), 0.64 (0.42–0.97), and 1.38 (0.75–2.54) in the DKD-non-Alb, DKD stage 3 Alb, DKD stages 1 to 2 Alb, and non-DKD subgroups, respectively (Pinteraction = 0.046; Supplementary Table S6). In this analysis, the non-DKD subgroup had the lowest event rate for time to first HHF.

Discussion

There are 4 major observations in this report. Our first major observation was the composition of the cohort with respect to the clinical DKD profile. Consistent with previous observations in people with T2DM, the proportion of participants with DKD-non-Alb was 10.8% in the overall VERTIS CV cohort. Among patients with baseline eGFR <60 ml/min per 1.73 m2 in VERTIS CV, 49.3% had DKD-non-Alb, similar to what has been described in other populations. Patients with eGFR <60 ml/min per 1.73 m2 at baseline in these analyses tended to be older, with a longer duration of T2DM, and requiring more insulin, lipid-modifying medications, and antihypertensive medications including diuretics, but with a lower use of metformin. Those with elevated albuminuria at baseline had anticipated clinical risk factors, such as higher SBP and HbA1c. Finally, for patients with DKD-non-Alb, previous studies have identified that females and individuals who had no history of tobacco use with improved glycemic control have increased preservation of normoalbuminuria, even with declining kidney function. In these analyses, a larger proportion of females and individuals with no history of tobacco use was observed in the DKD-non-Alb subgroup compared with the other subgroups.

The second observation was the acute increase in eGFR observed in placebo groups with baseline eGFR <60 ml/min per 1.73 m2. We have described this finding in previous analyses from VERTIS CV., The increase in eGFR in the placebo group in patients with CKD stage 3 has also been observed in the analyses of eGFR over time in other SGLT2 inhibitor cardiovascular outcomes trials.14, 15, 16 In addition, in a cohort of patients from an observational study using a Taiwanese database, an initial increase in eGFR was observed in patients taking glucose-lowering agents other than SGLT2 inhibitors. In the SGLT2 inhibitor cardiovascular outcomes trials, these subgroups of patients with CKD stage 3 had relatively low baseline UACR compared with patients from 2 SGLT2 inhibitor kidney outcomes studies who had elevated albuminuria levels at baseline, in which an increase in eGFR in the placebo groups was not observed., The basis for these eGFR changes in patients treated with placebo with CKD stage 3 and relatively low levels of albuminuria is not known, but it may be secondary to an improved diet (with lower sodium intake), which most often occurs when patients enter clinical studies, or regression to the mean.

In the current analysis, our third observation was that, in both placebo- and ertugliflozin-treated groups, chronic and total eGFR declines were slower in the DKD-non-Alb subgroup, followed by the non-DKD subgroup, with more rapid rates of eGFR decline in the DKD stages 1 to 2 Alb and DKD stage 3 Alb subgroups. We also observed a differential effect of eGFR decline in patients with similar baseline albuminuria levels whereby patients in the DKD stages 1 to 2 Alb subgroup had faster rates of eGFR decline compared with the DKD stage 3 Alb subgroup. Treatment with ertugliflozin compared with placebo resulted in significantly lower chronic (weeks 6 to 260) and total (weeks 0 to 260) eGFR slopes in all 4 subgroups. In a report from the EMPA-REG OUTCOME trial, the overall population was divided into the following 3 groups: patients with overt DKD (UACR >300 mg/g, regardless of eGFR), non-overt DKD (eGFR <60 ml/min per 1.73 m2 with UACR ≤300 mg/g), and all others (eGFR ≥60 ml/min per 1.73 m2 and UACR ≤300 mg/g). In the analysis from the EMPA-REG OUTCOME trial, chronic eGFR slope (week 4 to end of study) in patients randomized to placebo with overt DKD had a faster rate of eGFR decline than the non-overt DKD group (−6.00 and −0.74 ml/min per 1.73 m2 per year, respectively). In the VERTIS CV cohort, the comparable chronic eGFR slope (weeks 6 to 260) in patients randomized to placebo with macroalbuminuria at baseline (UACR >300 mg/g, regardless of baseline eGFR) was −3.3 ml/min per 1.73 m2 per year. For reasons that are not understood, the VERTIS CV, the placebo-treated population had a slower rate of eGFR decline compared with the EMPA-REG OUTCOME placebo-treated population.

Our fourth major observation was that, among patients randomized to placebo, the DKD-non-Alb subgroup had the lowest risk for the prespecified exploratory kidney composite outcome compared with the other groups, with an event rate of 0.3 per 100 person-years. The reduced risk for the prespecified exploratory kidney composite outcome in the DKD-non-Alb cohort is consistent with the finding that this group had the slowest rate of eGFR decline, as the prespecified exploratory kidney composite outcome is driven mainly by the sustained ≥40% decline from baseline in eGFR. In addition, these findings are also in agreement with results from the Chronic Renal Insufficiency study in patients with T2DM and reduced eGFR. In the Chronic Renal Insufficiency study, the absence of albuminuria was associated with a reduced risk for end-stage kidney disease, progression of CKD, and decline in eGFR slope. The reported effect of ertugliflozin on the kidney composite outcome was consistent across the studied groups (PInteraction = 0.34). For HHF end points, treatment with ertugliflozin was associated with a larger relative risk reduction in time to first HHF in patients with either elevated albuminuria and/or low eGFR at baseline compared with other kidney function subgroups, highlighting the larger heart failure benefits of ertugliflozin on patients with compromised kidney function who also have a high event rate for this outcome. Patients treated with placebo in the non-DKD subgroup had the lowest event rate for time to first HHF. The non-DKD subgroup had the lowest use of diuretics, including loop diuretics, at baseline compared with the other subgroups, irrespective of treatment randomization. These differences may have a role in the observed ertugliflozin treatment effect variation in the HHF outcome, in keeping with previous analyses from VERTIS CV.

This analysis does have limitations. First, it was a post hoc analysis and P values were not adjusted for multiplicity. Owing to the lack of multiplicity adjustment, these results are hypothesis generating. Finally, although there may be important differences between patients with microalbuminuria (UACR ≥30 and ≤300 mg/g) and macroalbuminuria (UACR >300 m/g) at baseline, especially within the DKD stage 3 Alb subgroup, further stratification analysis is unlikely to have sufficient power to detect differences.

In VERTIS CV, participants with DKD-non-Alb at baseline had the slowest rate of long-term eGFR decline and a lower risk for the prespecified exploratory kidney composite outcome. However, ertugliflozin was associated with reducing the risk of eGFR decline and the prespecified exploratory kidney composite outcome in all subgroups. The effect of ertugliflozin on the reduction in the risk for time to first HHF was larger in subgroups with more advanced kidney disease (patients with UACR ≥30 mg/g at baseline and/or eGFR <60 ml/min per 1.73 m2) and are resultantly at higher risk for this outcome.