Sodium–Glucose Cotransporter 2 Inhibition and Diabetic Kidney Disease

The impact of the current diabetes pandemic is rapidly approaching that of the Great Plague (1,2). Its prevalence has nearly quadrupled since the 1980s, and 1 in 10 adults, or 642 million people worldwide, are now projected to have diabetes by the year 2040 (3). As the number of people living with diabetes rises, the prevalence of diabetic complications is also rapidly escalating. Approximately half of individuals with type 2 diabetes (T2D) and one-third of people with type 1 diabetes (T1D) develop diabetic kidney disease (DKD), a microvascular complication that is now the leading cause of chronic kidney disease (CKD) and end-stage kidney disease (ESKD) in the world (4–6).

For people with diabetes, development of kidney disease increases the risk of death by five- to sixfold (7–9). Tragically, approximately 90% of patients with DKD die before requiring kidney replacement therapy (KRT). Among those who reach ESKD, the risk of death is 10- to 100-fold higher than for individuals with normal kidney function (10). Depending on the country, only 10%–50% of those who need KRT will ever receive it (10). Thus, in many parts of the world, ESKD equates to a virtual death sentence (10–12). Although survival rates for patients receiving KRT have improved modestly over the past few decades, the increased risk of death remains unacceptably high, as one-third of those treated by maintenance dialysis die within 3 years of initiation (13).

Achieving glycemic control with conventional blood glucose–lowering therapies early in the course of T1D or T2D reduces, but does not eliminate, the risk of developing DKD (11,14,15). Therefore, agents that control hyperglycemia safely while also preventing or treating DKD are urgently needed. Over the past three decades, discovery and elucidation of the role of sodium symporters in glucose reabsorption, and thereby glucose homeostasis, have pointed to sodium–glucose cotransporter 2 (SGLT2) inhibition as a viable therapeutic target (16–19). In cardiovascular disease (CVD) outcomes trials conducted for safety, SGLT2 inhibitors actually have demonstrated clear benefits on CVD and CKD. This Perspective highlights postulated mechanisms that may underlie clinical effects of SGLT2 inhibition and provides guidance for use of these antihyperglycemic agents in patients with T2D and CKD.

Under normoglycemic to mildly hyperglycemic conditions, the kidney reabsorbs almost all glucose in the glomerular filtrate (19). Glucose reabsorption occurs against its concentration gradient and is driven by sodium symporters expressed in the proximal tubule (20). Of these, SGLT2 and sodium–glucose cotransporter 1 (SGLT1) are the principal known contributors. The complementary glucose transport kinetics of these two transporters permit almost complete resorption of filtered glucose (21). Experimental data indicate that SGLT2 is expressed on the luminal surface of the epithelial cells of the proximal convoluted tubule and is a low-capacity, high-affinity glucose transporter (K_m ∼1–4 mmol/L for glucose) with 1:1 Na⁺/glucose stoichiometry. As such, SGLT2 is responsible for the reabsorption of ∼90% of filtered glucose (Fig. 1). SGLT1 is expressed on the luminal surface of the epithelial cells of the late proximal tubule and reabsorbs most of the remaining ∼10% of filtered glucose (21–23).

In humans, glycosuria occurs when blood glucose reaches a threshold of about 180 mg/dL (10 mmol/L). However, this threshold can range approximately 100–240 mg/dL (5.5–13 mmol/L) (24–27). Diabetes increases the glycosuric threshold to 200–240 mg/dL (11–13 mmol/L) and, in this way, exacerbates hyperglycemia. The exact mechanism behind this response is unclear but most likely includes increased expression of SGLTs. In studies of mouse and rat models of T2D, SGLT1 and SGLT2 expression are increased in the diabetic kidney (28–30). Correspondingly, tubular epithelial cells freshly isolated from the urine of humans with T2D exhibit increased expression of SGLT2, and kidney tissue from patients with T2D displays higher expression of SGLT1 protein and mRNA (31, 32). In sum, higher glucose reabsorptive capacity of the diabetic kidney likely results from increased expression of SGLTs (33) (Fig. 1).

The BI 10773 (Empagliflozin) Cardiovascular Outcome Event Trial in Type 2 Diabetes Mellitus Patients (EMPA-REG OUTCOME) trial and the Canagliflozin Cardiovascular Assessment Study (CANVAS) Program were the original large studies that demonstrated improvements in both CVD and CKD outcomes in over 17,000 participants with T2D at high CVD risk (34–38). EMPA-REG OUTCOME enrolled approximately 7,000 participants and followed them for a mean duration of 3.1 years. Study participants were randomized to empagliflozin (10 mg or 25 mg) or placebo. The empagliflozin group experienced significantly lower rates of hospitalization for heart failure (35% relative risk reduction), death from CVD (38% relative risk reduction), and death from any cause (32% relative risk reduction) compared with the placebo group (34). Importantly, these observed risk reductions were maintained across estimated glomerular filtration rate (eGFR) and albuminuria categories in more than 2,000 participants with eGFR <60 mL/min/1.73 m² and/or macroalbuminuria (35).

EMPA-REG OUTCOME also examined secondary kidney disease outcomes of incident or worsening nephropathy: new-onset albuminuria or progression to urine albumin-to-creatinine ratio (UACR) >300 mg/g (macroalbuminuria), doubling of serum creatinine, initiation of KRT, and death from kidney disease as a composite outcome and individual outcomes (36). The relative risk of developing incident or worsening nephropathy was 39% lower in the empagliflozin group compared with placebo (13% vs. 19%, P < 0.001) (36). A comparable relative risk reduction for nephropathy was observed in participants with CVD who underwent coronary artery bypass graft surgery (37). Notably, most participants in EMPA-REG OUTCOME also received treatment with ACE inhibitors or angiotensin receptor blockers, agents that have been shown to reduce DKD progression and prevent ESKD.

The CANVAS Program integrated data from two CVD outcome trials enrolling over 10,000 participants with T2D, randomized to either canagliflozin or placebo and followed for a mean duration of 3.6 years (38). The primary composite outcome of death from CVD causes, nonfatal myocardial infarction, and nonfatal stroke occurred at a significantly lower rate in the canagliflozin group compared with placebo (14% relative risk reduction, P < 0.001). The risk of progression to albuminuria was decreased by 27%, and the composite kidney disease outcome (40% eGFR decline, KRT, or death from kidney causes) occurred 40% less frequently in the canagliflozin group relative to placebo (38). The secondary analysis of the CANVAS Program showed that cardiovascular and kidney outcomes were consistent across the different levels of kidney function (eGFR 30–45, 45–60, 60–90, and ≥90 mL/min/1.73 m²); however, canagliflozin treatment had greater benefits on fatal/nonfatal strokes in groups with eGFR <60 mL/min/1.73 m² (hazard ratio compared with placebo was 0.56 in the 45–60 mL/min/1.73 m² group and 0.32 in the 30–45 mL/min/1.73 m² group) (39). Ongoing cardiovascular outcomes trials with two other SGLT2 inhibitors, dapagliflozin and ertugliflozin, will also report major CKD outcomes (40).

Although the findings from the EMPA-REG OUTCOME and CANVAS trials provide a strong signal that SGLT2 inhibition preserves kidney function and improves overall and kidney survival in T2D, results from two clinical trials primarily designed to evaluate CKD outcomes with SGLT2 inhibition are keenly awaited. The Canagliflozin and Renal Endpoints in Diabetes with Established Nephropathy Clinical Evaluation (CREDENCE) (ClinicalTrials.org identifier NCT02065791) and A Study to Evaluate the Effect of Dapagliflozin on Renal Outcomes and Cardiovascular Mortality in Patients With CKD (Dapa-CKD) (ClinicalTrials.org identifier NCT03036150) trials are evaluating effects of canagliflozin or dapagliflozin on composite primary outcomes including ESKD, doubling of serum creatinine (CREDENCE), ≥50% sustained decline in eGFR (Dapa-CKD), and kidney disease or CVD death in participants with established DKD (41). CREDENCE concluded early due to positive efficacy findings, and results are expected to be publicly released in early 2019 (42). Dapa-CKD is expected to report in 2021 (Table 1).

Preservation of eGFR and albuminuria reduction are class effects of SGLT2 inhibitors. An initial effect observed within the first few weeks of SGLT2 inhibition is the reduction of eGFR by approximately 5 mL/min/1.73 m², followed by stabilization over time (36,43–49). This phenomenon has been observed in patients with eGFR as low as 30 mL/min/1.73 m² (50). Compared with glimepiride, canagliflozin resulted in slower mean eGFR decline (0.5 mL/min/1.73 m² per year for canagliflozin 100 mg daily, 0.9 mL/min/1.73 m² per year for canagliflozin 300 mg daily, and 3.3 mL/min/1.73 m² per year with glimepiride, P < 0.01 for between-group comparisons), despite achieving a similar level of glycemic control (46). Albuminuria reduction is also observed across levels of albuminuria and eGFR. Although it did not have a significant effect on development of new-onset albuminuria in EMPA-REG OUTCOME, empagliflozin produced a 38% relative risk reduction in progression to severely increased albuminuria (11% vs. 16%, P < 0.001) compared with placebo (36). Similarly, in the CANVAS Program, canagliflozin produced a 27% reduction in progression to severely increased albuminuria and 1.7-fold higher rate of albuminuria regression (38). Among patients with baseline UACR >100 mg/g, treatment with dapagliflozin decreased 24-h urine albumin excretion by 36% (P < 0.001), and among those with eGFR 30–60 mL/min/1.73 m², it decreased frequency of severely increased albuminuria (UACR >1,800 mg/g) compared with placebo (44, 51). Among patients with eGFR ≥30 to <50 mL/min/1.73 m², treatment with canagliflozin was associated with greater decrease in UACR compared with placebo (median percent reduction −30%, −21%, and −8% in canagliflozin 100 mg daily, 300 mg daily, and placebo groups, respectively) (48). When compared with glimepiride treatment with canagliflozin 100 mg or 300 mg daily in patients with at least moderately increased albuminuria (UACR ≥ 30 mg/g), decreased UACR by 32% (P = 0.01) and 50% (P < 0.001), respectively, despite similar glycemic control (46).

Knowledge of the biological mechanisms behind the kidney-protective effects of SGLT2 inhibition is evolving. Although blood glucose lowering is central to DKD prevention, there are also likely direct effects independent of glycemia.

One putative mechanism is normalization of glomerular hemodynamics through restoration of tubuloglomerular feedback. Hyperfiltration with resulting hypertension in the glomerular capillary circulation is an early hemodynamic change observed in at least 75% of patients with T1D and 40% of those with T2D (Fig. 1) (52,53). Glomerular hyperfiltration is driven by metabolic derangements including hyperglycemia and hyperaminoacidemia, as well as increased proximal tubular reabsorption of glucose and sodium chloride via SGLT1 and SGLT2 (Fig. 2).

Tubuloglomerular feedback is an adaptive mechanism through which reabsorption of sodium and chloride in the macula densa promotes adenosine release (Fig. 2). Adenosine, in turn, acts in paracrine manner to constrict the afferent arteriole. In diabetes, as a result of increased reabsorption of sodium and chloride in the proximal tubule, delivery to the macula densa is decreased, leading to lower solute reabsorption and a consequent decrease in adenosine production. By promoting relative afferent arteriolar vasodilation, this mechanism contributes to glomerular hyperperfusion, hypertension, and hyperfiltration in diabetes (54).

By blocking reabsorption of sodium chloride in the proximal tubule, SGLT2 inhibition restores solute delivery to the macula densa and thereby restores normal tubuloglomerular feedback (Fig. 2). A net effect is reversal of afferent vasodilation and normalization of glomerular hemodynamics (55). This effect has been observed with the nonspecific SGLT2 inhibitor phlorizin in a T1D model in rats and, more recently, with the selective SGLT2 inhibitor empagliflozin in a mouse T1D model (56,57). In humans with T1D and glomerular hyperfiltration, treatment with empagliflozin decreased directly measured GFR (inulin clearance) by 33 mL/min/1.73 m² (mean ± SD 172 ± 23 mL/min/1.73 m² to 139 ± 25 mL/min/1.73 m²) in conjunction with decreased plasma flow to the kidney, lower plasma nitric oxide levels, and increased kidney vascular resistance. This effect was only observed in patients with diabetes with glomerular hyperfiltration (58).

SGLT2 inhibition may have additional anti-inflammatory and antifibrotic actions that protect the kidney. In primary proximal tubular cells, SGLT2 inhibition suppressed the generation of a hyperglycemia-mediated increase in reactive oxygen species (47,59). Experimental rat and mouse models of diabetes have shown attenuation of glomerulosclerosis and tubulointerstitial fibrosis with SGLT2 inhibition (60–62). Decreased urinary excretion of markers of kidney tubular injury (e.g., kidney injury molecule 1) and inflammatory markers (e.g., interleukin-6) have been observed in humans with T2D treated with dapagliflozin (47).

Glycemic control is known to decrease risk of DKD onset, particularly if implemented early in the course of diabetes (14,15). In patients with diabetes and preserved kidney function, SGLT2 inhibition reduces HbA_1c by approximately 1% (63). Due to the intrinsic mechanism of action, the glycemic-lowering effects of SGLT2 inhibitors are blunted in patients with low eGFR (36,39,43,44,48,63,64). For instance, the adjusted mean treatment difference in HbA_1c was −0.7% (P < 0.001) in patients with eGFR >60 and ≤90 mL/min/1.73 m² who received empagliflozin when compared with placebo, and in those with eGFR >30 and ≤60 mL/min/1.73 m², the adjusted mean difference was −0.4% (P < 0.001) (43). Pooled analysis of phase III empagliflozin clinical trials confirmed this finding with evidence of placebo-corrected reductions in HbA_1c decreasing with declining eGFR (64). Treatment with dapagliflozin reduced HbA_1c between 0.3% and 0.4% in patients with eGFR >45 and ≤60 mL/min/1.73 m². No HbA_1c reduction was observed in patients with eGFR ≤40 mL/min/1.73 m² (44). As such, the antihyperglycemic effects of SGLT2 inhibition seem less likely to confer kidney protection in the setting of moderate-to-severe CKD.

As body fat loss per se may decrease albuminuria and glomerular hyperfiltration, weight reduction effect of SGLT2 inhibition may indirectly protect the diabetic kidney (44,55). In patients with normal kidney function, SGLT2 inhibition leads to a loss of 60–80 g of glucose (240–320 calories) per day via glycosuria, with expected weight loss of 2–3 lb (0.9–1.4 kg) per month (65). However, weight loss plateaus after about 6 months of treatment, after achieving a total weight loss of 5–7 lb (2.3–3.2 kg) (63). After more than 2 years of dapagliflozin treatment in patients with T2D and a mean weight of 225 lb (102 kg), experienced weight loss was 11 lb (5 kg) with a concomitant decrease in waist circumference (66,67). Notably, a recent pooled analysis of phase III empagliflozin trials and secondary analysis of the CANVAS Program found that the weight loss effects were maintained in patients with eGFR as low as 30 mL/min/1.73 m² (39, 64).

Antihypertensive effects are observed with empagliflozin, dapagliflozin, and canagliflozin. Each of them lower systolic blood pressure by approximately 5 mmHg and diastolic blood pressure by approximately 2 mmHg (63,68–70). The systolic blood pressure reduction appears greatest within 3–4 months of initiation of treatment with empagliflozin and dapagliflozin (34,66). In contrast to blood glucose lowering, the magnitude of blood pressure reduction is maintained, or perhaps increased, in patients with low eGFR (39). For example, in patients with T2D, the mean placebo-corrected changes in systolic blood pressure among those treated with empagliflozin were −3 mmHg with eGFR ≥90 mL/min/1.73 m², −4 mmHg with eGFR 60–89 mL/min/1.73 m², −6 mmHg with eGFR 30–59 mL/min/1.73 m², and −7 mmHg with eGFR <30 mL/min/1.73 m² (64). The mechanisms underlying blood pressure reduction are likely multiple and may include natriuresis, weight loss, and improved endothelial function and vascular compliance (71–76).

The natriuretic effect may be enhanced in diabetes due to greater proximal tubular sodium reabsorption related to increased expression of SGLT2 and SGLT1 (77, 78). Another postulated mechanism for the natriuretic effect of SGLT2 inhibition is “cross talk” with other solute transporters, including the Na⁺/H⁺ exchanger 3 (NHE3). NHE3 is responsible for much of the sodium reabsorption from the glomerular filtrate (79). In rats, SGLT2 and NHE3 colocalize in the membrane of proximal tubular cells (80). SGLT2 inhibition with phlorizin inhibits sodium bicarbonate reabsorption by NHE3, though the specific mechanism of this effect remains unclear (81).

Since the U.S. Food and Drug Administration (FDA) approval of canagliflozin for the treatment of T2D in 2013, the SGLT2 inhibitor class has quickly gained usage. Major guidelines and consensus statements, such as the American Diabetes Association (ADA) Standards of Medical Care in Diabetes and the American Association of Clinical Endocrinologists (AACE)/American College of Endocrinology (ACE) algorithm for the comprehensive management of people with T2D, recommend SGLT2 inhibition because of the combined effects on glycemia, weight, and blood pressure in people with preserved eGFR (82–84). Based largely on results of the EMPA-REG OUTCOME and CANVAS clinical trials, the consensus report from the ADA and European Association for the study of Diabetes (EASD) recommends use of SGLT2 inhibitors as an add-on antihyperglycemic therapy of choice in patients who have CVD or CKD (84). Dosing recommendations (eGFR >30 mL/min/1.73 m² for dapagliflozin, canagliflozin, and ertugliflozin and >45 mL/min/1.73 m² for empagliflozin), which are based on the limited antihyperglycemic efficacy of SGLT2 inhibition in patients with lower eGFR, are not changed.

Though both empagliflozin and canagliflozin have now been approved by the FDA for the indication of reducing the risk of cardiovascular events and cardiovascular death in adults with T2D and established CVD, to date, SGLT2 inhibitors have not been recommended for the express purpose of improving CKD outcomes (85,86). The current recommendations to limit use of SGLT2 inhibitors by eGFR criteria may change once results of CREDENCE and other ongoing clinical trials with primary CKD outcomes are reported (Table 2) (41,42,87).

SGLT2 inhibitors show great promise for prevention and treatment of DKD. Trials with empagliflozin have demonstrated, for the first time, a reduction in all-cause and cardiovascular mortality in patients with T2D and CKD. The mortality risk in this population has heretofore been unacceptably high and largely unmitigated; thus, the importance of improving survival while maintaining kidney function in patients with DKD is of urgent and utmost importance. Research is needed to inform the use of SGLT2 inhibitors in the setting of T1D and perhaps for indications outside of diabetes, such as CKD without diabetes. Progress on these fronts is already under way. For example, the dual SGLT1/2 inhibitor sotagliflozin is currently under study for use in patients with T1D (88,89). Empagliflozin will soon be studied for primary CKD outcomes and cardiovascular deaths among those with established diabetic and nondiabetic CKD. Elucidation of the biological mechanisms underlying the effects of SGLT2 inhibition is necessary to advance understanding and more fully optimize clinical applications of these agents for the treatment of diabetes, CKD, and CVD.