Sodium–Glucose Cotransporter 2 Inhibitors and the Kidney

The natural history of DKD is shown in Figure 1 (9,10). Hyperfiltration (increased GFR), increased intraglomerular pressure, and glomerular hypertrophy (9–11) are present in the early stages of DKD in many, but not all, patients with diabetes and in some, but not all, studies have been found to predict the progression to late-stage renal disease in both type 1 diabetes (12,13) and type 2 diabetes (14). In a large, prospective study involving 1,388 kidney donors, a high single-nephron GFR (resulting from intraglomerular hypertension) was shown to be associated with biopsy-proven large glomeruli and glomerulosclerosis (15), providing support for glomerular hypertension and hypertrophy in the pathogenesis of DKD.

Despite the development of glomerulosclerosis, basement membrane thickening, increased mesangial matrix material, and tubule-interstitial disease, there are no clinical or laboratory clues that predict who is at risk for the development of DKD. In most individuals destined to develop DKD, the first laboratory evidence of renal disease is the development of microalbuminuria (30–300 mg/dL), followed ∼5 years later by the onset of overt albuminuria (>300 mg/day) (Figure 1). Within 5 years of the onset of macroalbuminuria, serum creatinine will have doubled, and within another 4–5 years, individuals will have progressed to ESRD and require dialysis or transplantation.

Two caveats need to be emphasized about albuminuria. First, in type 2 diabetes, microalbuminuria is a much stronger predictor of ASCVD than it is of DKD (16,17). Second, ∼20% of patients with diabetes progress to ESRD without albuminuria (18). Thus, urine albumin excretion lacks both the sensitivity and the specificity to detect early DKD, and novel markers to identify progressors are needed.

Multiple factors have been implicated in the development and progression of DKD (Table 1) (19). Hyperglycemia is the most important factor, and, if the A1C remains within the normal range (≤5.7%), DKD does not occur. Microalbuminuria is present in ∼15–20% of patients with prediabetes (A1C 5.8–6.4%) (20), but it is unknown how many of these individuals progress to advanced DKD as long as their A1C remains in the prediabetic range. When A1C rises above 7.0%, there is a gradual but increasingly steep rise in the incidence of microalbuminuria (21). The importance of good glycemic control in preventing diabetic microvascular complications has been underscored by the landmark Diabetes Control and Complications Trial (22) in type 1 diabetes, the UK Prospective Diabetes Study (23) in type 2 diabetes, and many subsequent studies.

Hypertension is the second most important factor that accelerates the progression of DKD (24–26). The American Diabetes Association and the Eighth Joint National Committee recommend a blood pressure <140/90 mmHg in most people with diabetes and <130/80 mmHg in those at high risk for ASCVD and renal disease, whereas the American Heart Association/American College of Cardiology recommend a blood pressure <130/80 mmHg in all individuals with diabetes, with or without kidney disease (27).

Deranged tubuloglomerular feedback plays a central role in the pathogenesis of DKD (28–31). In people with either type 1 or type 2 diabetes, the filtered load of glucose is increased, leading to enhanced glucose along with sodium reabsorption by the sodium–glucose cotransporter 2 (SGLT2) protein in the proximal tubule. This process leads to a reduction in delivery to and absorption of sodium chloride by the macula densa cells. This reduction is perceived by the kidney as a decrease in circulating vascular volume and leads to 1) activation of the local renin-angiotensin system, resulting in efferent arteriolar vasoconstriction, glomerular hypertension, and renal hyperfiltration; and 2) decreased production of adenosine (a potent vasoconstrictor), leading to afferent arteriolar vasodilation, enhanced renal plasma flow, increased intraglomerular pressure, and hyperfiltration (Figure 2). On a long-term basis, the increase in intraglomerular pressure promotes glomerular sclerosis, albuminuria, and decline in GFR.

In animal models of diabetic nephropathy, tubular hypertrophy precedes the development of glomerular hypertrophy, and inhibition of tubular hypertrophy with ornithine decarboxylase prevents the development of glomerular hypertrophy, glomerular hypertension, glomerulosclerosis, and DKD (32). This observation suggests that tubular hypertrophy, stimulated by increased proximal tubular glucose-sodium reabsorption and/or one of the multiple growth factors (e.g., angiotensin, transforming growth factor-β, insulin-like growth factor 1, platelet-derived growth factor, and vascular endothelial growth factor) that are released by the diabetic kidney, precedes and contributes to the glomerular hypertrophy, which, once established, increases the intraglomerular pressure according to Laplace’s law.

Renal hypoxia has been demonstrated in both human and animal models of diabetes and has been implicated in the development and progression of DKD (33–36). In the diabetic kidney, the increase in renal plasma flow/GFR results in an increased filtered load of sodium and enhanced tubular sodium reabsorption, which requires adenosine triphosphate (ATP) production via the mitochondrial electron transport chain, a process that requires oxygen. Because renal blood flow, and therefore oxygen delivery, to the kidney is limited, a mismatch between supply and demand occurs, leading to renal hypoxia, which promotes a fibrotic response and diabetic nephropathy. The presence of renal hypoxia has been demonstrated using blood oxygenation level–dependent magnetic resonance imaging in humans and oxygen microelectrodes in diabetic animal models as well as by increased renal expression of hypoxia-inducible factor 1α (37).

The podocytes are an integral component of the glomerular filtration barrier. In DKD, podocyte number and attachment to the glomerular basement membrane are reduced (38), allowing albumin to escape into the glomerular filtrate. Albumin is toxic to the kidney and promotes glomerulosclerosis (39). Consistent with this, each 30% decrease in albuminuria reduces the risk of ESRD by ∼27% (40).

“Fatty kidney disease,” or obesity-related glomerulopathy, is a well-established cause of renal injury (41,42). Fatty kidney disease is diagnosed when no other primary nephropathy is evident and is characterized by glomerulomegaly with glomerulosclerosis, mesangial cell proliferation, podocyte loss, tubular hypertrophy, and glomerular hyperfiltration. Lipid deposition is demonstrable within podocytes, mesangial cells, and tubular cells (41–45), and many genes involved in lipid metabolism are differentially expressed (46). Thus, fatty kidney disease closely resembles the histologic changes observed in DKD. Of note, most patients with type 2 diabetes are obese, making it difficult to distinguish between the contribution of lipotoxicity and that of the diabetic state per se to the development of DKD.

Inflammation, increased reactive oxygen species, and mitochondrial dysfunction (47,48) also have been shown to play a role in the development of DKD, but these are likely to be late events triggered by the disturbances described above.

Finally, genetics plays an important role in the development of DKD, as demonstrated by the familial and ethnic clustering of the disease (49). Genome-wide association studies have identified >100 genetic variants associated with DKD in both type 1 and type 2 diabetes (50), but causality has yet to be established.

DKD does not occur in the absence of hyperglycemia; therefore, maintaining an A1C <6.5% is the most important treatment for primary prevention of DKD. In patients with established DKD, comprehensive treatment involves prevention of progressive renal disease, macrovascular complications, and further microvascular complications.

After hyperglycemia, hypertension is the single most important factor that accelerates the rate of progression of DKD (24–26). Therefore, maintaining tight blood pressure control—<130/80 mmHg—is essential for slowing the development and progression of DKD (27). A high-protein diet has been shown to accelerate the progression of kidney disease in people with or without diabetes (51), and amino acid infusion increases renal plasma flow and GFR (52) because of an increase in intraglomerular pressure. Therefore, ingestion of a diet high in protein should be avoided. In a large, prospective study, a low-protein diet in patients treated with an ACE inhibitor failed to slow the rate of progression of CKD (53). Therefore, a normal protein intake is recommended in patients with diabetes and renal impairment. Because obesity is associated with lipid deposition within the kidney and glomerulosclerosis (41–44), weight loss should be encouraged (54). Of course, these lifestyle interventions should be instituted in all patients with diabetes, along with strict glycemic control, to prevent the development of DKD.

Until recently, ACE inhibitors and angiotensin receptor blockers (ARBs) have been the pharmacologic mainstay for slowing the progression of DKD (55–57). As discussed previously, activation of the renin-angiotensin system plays a pivotal role in the development and progression of DKD by 1) causing vasoconstriction of the efferent arteriole, with a resultant increase in intraglomerular pressure (Figure 1); 2) its growth-promoting properties, leading to glomerular hypertrophy; and 3) activating proinflammatory and profibrotic pathways. Although inhibition of the renin-angiotensin-aldosterone system (RAAS) reduces hyperfiltration, it does not completely normalize it (58) and does not completely prevent kidney injury (59). Therefore, additional therapeutic interventions are required to stop the progression of DKD. As detailed in Table 1, multiple factors contribute to the development and progression of DKD. It follows, then, that multiple pharmacologic therapies will be required to slow the progression of DKD.

Mineralocorticoid receptor antagonists (MRAs), including spironolactone and eplerenone, reduce albuminuria when conjointly administered with an RAAS blocker in patients with diabetes (60,61), but hyperkalemia has tempered enthusiasm for this combination therapy. In the recent FIDELIO-DKD (Finerenone in Reducing Kidney Failure and Disease Progression in Diabetic Kidney Disease) trial (62), the nonsteroidal MRA finerenone was shown to reduce the combined primary end point of CKD progression, kidney failure, or kidney death when added to standard care in patients with chronic DKD. Although not yet approved by the U.S. Food and Drug Administration, this agent, added to ACE inhibitor or ARB therapy, may be an effective combination therapy for retarding the progression of DKD. However, monitoring for hyperkalemia still would be prudent. Although uric acid has been suggested to play a role in the progression of CKD, the recent PERL (Preventing Early Renal Loss in Diabetes) study (63) failed to demonstrate any benefit of uric acid lowering in slowing the rate of decline in GFR.

SGLT2 inhibitors have multiple nonrenal benefits that have been reviewed in detail and will be briefly covered here (Figure 3). Currently, there are four approved SGLT2 inhibitors in the United States: empagliflozin, dapagliflozin, canagliflozin, and ertugliflozin.

The primary mechanism of action of all SGLT2 inhibitors is inhibition of glucose/sodium reabsorption by the SGLT2 transporter in the proximal tubule (64). In type 2 diabetes, the normal renal threshold (180 mg/dL) for glucose spillage into the urine is increased and rises progressively with increasing A1C. Treatment of patients with type 2 diabetes, as well as those with normal glucose tolerance, with an SGLT2 inhibitor decreases the renal threshold for glucose spillage into urine to ∼40 mg/dL, leading to the excretion of ∼70–80 g glucose per day (65,66). Because the renal threshold is reduced to below the fasting plasma glucose concentration (∼80–90 mg/dL) in normal glucose-tolerant subjects, even patients without diabetes respond with a glucosuric response of ∼50–60 g/day. However, hypoglycemia does not occur because the liver reads the amount of glucose lost in the urine and augments its production of glucose to quantitatively match the amount excreted in the urine. Consequently, in patients with type 2 diabetes who are treated with an SGLT2 inhibitor, hypoglycemia does not occur (64,65). Of note, the weight loss and blood pressure–lowering effects of SGLT2 inhibitors are maintained even in individuals with diabetes who have markedly reduced GFR. Furthermore, because of the marked reduction in the renal threshold, there is no such thing as a nonresponder, as long as renal function remains relatively intact (GFR >60 mL/min/1.73 m²). Additionally, because the primary mechanism of action of the SGLT2 inhibitors is on the kidney, they can be used in combination with all other antidiabetic agents to further reduce A1C (63,64).

SGLT2 inhibition reduces both fasting and postprandial plasma glucose concentrations, leading to the reversal of glucotoxicity. This process results in a marked increase in insulin secretion (67,68) and a modest improvement in insulin sensitivity (69). Because calories are lost in the urine, weight loss in the amount of 2–3 kg occurs within the initial 3–6 months of SGLT2 inhibitor therapy (63,64). Because SGLT2 inhibitors block sodium along with glucose absorption, a negative sodium balance of ∼100 mEq and a negative fluid balance of ∼750 cc is observed within the first 48 hours (66), leading to a decrease in systolic/diastolic blood pressure of 4–5/1–2 mmHg (63,64).

There have been four major cardiovascular outcomes trials (CVOTs) of SGLT2 inhibitors: EMPA-REG OUTCOME (BI 10773 [Empagliflozin] Cardiovascular Outcome Event Trial in Type 2 Diabetes Mellitus Patients) (70), the CANVAS (Canagliflozin Cardiovascular Assessment Study) Program (71), DECLARE-TIMI 58 (Dapagliflozin Effect on Cardiovascular Events) (72), and VERTIS CV (Randomized, Double-Blind, Placebo-Controlled, Parallel-Group Study to Assess Cardiovascular Outcomes Following Treatment With Ertugliflozin [MK-8835/PF-04971729] in Subjects With Type 2 Diabetes Mellitus and Established Vascular Disease) (73). In all four trials (70–73), a composite of major adverse cardiovascular events (MACE) was the primary end point. Their results have been summarized in two recent meta-analyses (74,75).

In two CKD trials, CREDENCE (Evaluation of the Effects of Canagliflozin on Renal and Cardiovascular Outcomes in Participants With Diabetic Nephropathy) (76) and DAPA-CKD (A Study to Evaluate the Effect of Dapagliflozin on Renal Outcomes and Cardiovascular Mortality in Patients With Chronic Kidney Disease) (77), MACE was a secondary outcome.

In the combined analysis (71,75), MACE was reduced by 12%, cardiovascular death by 17%, and myocardial infarction by 12%. There was no significant reduction in stroke risk in any of the four trials (hazard ratio [HR] 0.96, 95% CI 0.86–1.09). Hospitalization for heart failure (HHF) was reduced by 32%, and the composite end point of cardiovascular death and HHF was reduced by 24%. Heterogeneity for cardiovascular mortality was noted, primarily resulting from the large 38% reduction in cardiovascular death in the EMPA-REG OUTCOME trial. However, it should be noted that 99% of the patients with type 2 diabetes in that trial (70) had a prior cardiovascular event, compared with ∼66% in the CANVAS Program (71) and ∼40% in the DECLARE-TIMI 58 trial (72). Thus, the heterogeneity for cardiovascular mortality most likely was secondary to differences in the patient population (i.e., the percentage of patients with a previous cardiovascular event versus the percentage of patients with cardiovascular risk factors but without a prior event) than to intrinsic differences between the three SGLT2 inhibitors.

In the VERTIS CV trial (73), the MACE end point was not significantly different from placebo (HR 0.97). In the recently published DAPA-HF (Study to Evaluate the Effect of Dapagliflozin on the Incidence of Worsening Heart Failure or Cardiovascular Death in Patients With Chronic Heart Failure) trial (78), patients with and without diabetes who had heart failure and reduced ejection fraction experienced a 35% decrease in the primary end point of worsening heart failure or cardiovascular death. Similar findings have been reported with empagliflozin in the EMPEROR-Reduced (Cardiovascular and Renal Outcomes With Empagliflozin in Heart Failure) trial (79).

Collectively, these studies provide conclusive evidence that SGLT2 inhibitors provide protection against cardiovascular death, myocardial infarction, and HHF in patients with type 2 diabetes who have had a prior cardiovascular event (70–73), as well as in those with type 2 diabetes who have risk factors for ASCVD (72). The results of these CVOTs and potential mechanisms for the cardioprotective effects of SGLT2 inhibitors are discussed in more detail in other articles in this Diabetes Spectrum From Research to Practice section (pp. 214–256) and elsewhere (78,80).

The effect of SGLT2 inhibitors on renal function was a secondary outcome in the EMPA-REG OUTCOME trial (70), the CANVAS Program (71), and the DECLARE-TIMI 58 trial (72). A composite renal outcome including doubling of the serum creatinine or >40% sustained reduction in GFR, renal replacement therapy (dialysis or transplantation), and renal death was examined in all three studies and found to be significantly reduced. Furthermore, benefit was demonstrated for each component of the composite outcome individually. In a meta-analysis (75) including all three CVOTs, the renal composite outcome was reduced by 38% (HR 0.62, 95% CI 0.56–0.70, P <0.001) with no heterogeneity among the studies. Although these results are impressive, these studies mainly included subjects with normal to mildly impaired renal function; only a minority (10–25%) had a baseline estimated GFR (eGFR) <60 mL/min/1.73 m².

The CREDENCE trial (74), which included people with type 2 diabetes and an eGFR of 30–90 mL/min/1.73 m² or urine albumin-to-creatinine ratio (UACR) of 300–5,000 mg/g, has provided definitive evidence about the renal protective effect of SGLT2 inhibitors. Canagliflozin versus placebo reduced the renal composite outcome (doubling of serum creatinine; sustained decrease in GFR to <15 mL/min/1.73 m² for ≥30 days; dialysis for ≥30 days or renal transplant; or renal death) by 34% (HR 0.66, 95% CI 0.53–0.81, P <0.001) over a follow-up period of 2.6 years (Figure 4). Furthermore, each component of the primary outcome was significantly reduced individually. When stratified by eGFR or UACR, subjects with an eGFR <60 mL/min/1.73 m² or UACR >1,000 had the greatest renal benefit. Among 1,000 patients treated for 1.5 years, 22 would need to be treated with canagliflozin to prevent the renal-specific composite outcome.

The results of the DAPA-CKD trial (75) were similar to those of the CREDENCE trial. The renal composite end point (sustained GFR decline >50%, ESRD, or renal death) was decreased by 44% (HR 0.56, 95% CI 0.45–0.68, P <0.001) over a follow-up period of 2.4 years (Figure 4). Each component of the renal composite end point was significantly reduced individually, and patients with or without diabetes benefited similarly.

Of great importance, all patients in the CREDENCE and DAPA-CKD trials were required to be on a stable dose of ACE inhibitor or ARB for at least 4 weeks before randomization. Thus, the renal benefit of canagliflozin and dapagliflozin was in addition to that provided by an ACE inhibitor or ARB. Few patients in these trials had an eGFR <25–30 mL/min/1.73 m². However, there is no reason to believe that the benefit observed in patients with type 2 diabetes and an eGFR of 30–45 mL/min/1.73 m² would be any different from that in individuals with an eGFR <30 mL/min/1.73 m², nor would one expect the side effect profile to be any more adversely affected. Therefore, the authors believe that SGLT2 inhibitor therapy should be maintained until the time a patient is ready for renal replacement therapy. The similar results in the CREDENCE and DAPA-CKD trials support the concept that the renal-protective effect of the SGLT2 inhibitors is a class effect.

Now that two classes of drugs, SGLT2 inhibitors and ACE inhibitors/ARBs, are available for the treatment of DKD (note that these two classes are also cardioprotective), clinicians must consider whether both drugs should be started simultaneously at the time of DKD diagnosis or whether they should be added sequentially. Because of the ominous prognosis for patients with DKD, the authors favor starting both medications simultaneously or within a 2- to 3-week interval until a large prospective study can directly address this question. Another unanswered question is at what level of renal impairment (i.e., microalbuminuria versus macroalbuminuria or eGFR <60 versus >60 mL/min/1.73 m²) should an SGLT2 inhibitor be started to derive the maximum benefit? The authors favor the earlier start, recognizing that the predictive value of microalbuminuria for future development of ESRD is much less than that of macroalbuminuria and that ∼20% of patients with diabetes and DKD do not manifest macroalbuminuria.

The mechanisms that contribute to the development of DKD were reviewed above and are the subject of several recent reviews (11,19,81). In patients with type 2 diabetes and an eGFR <45 mL/min/1.73 m², the glucosuric effect and, therefore, the glucose-lowering effect of SGLT2 inhibitors is markedly attenuated, making improved glycemic control an unlikely explanation for their renal protective effect. Furthermore, it remains to be determined at what level of albuminuria intensive glycemic control fails to retard the progression of established DKD. When high levels of macroalbuminuria (>0.5–1.0 g/day) are present, improved glucose control is unlikely to halt the progression of DKD (82,83), although it may be able to slow progression down (84,85).

Strict glycemic control (A1C <6.5%) is of paramount importance in preventing the onset of DKD (i.e., primary prevention). To the extent that SGLT2 inhibitor therapy helps to normalize A1C in patients with newly diagnosed type 2 diabetes without evident renal disease, it can help prevent the development of DKD.

Hypertension is a key factor that accelerates the progression of DKD. SGLT2 inhibitors routinely reduce systolic/diastolic blood pressure by 4–5/1–2 mmHg (64) and thus could contribute to slowing the progression of DKD. However, the reduction in blood pressure is modest, and factors in addition to blood pressure reduction must play a more central role in the renal protective effect of SGLT2 inhibitors observed in the EMPA-REG OUTCOME, CANVAS Program, DECLARE-TIMI 58, CREDENCE, and DAPA-CKD trials (70–74).

Increased intraglomerular pressure is a well-established pathophysiologic factor in the development of diabetic nephropathy (28–31). SGLT2 inhibitors, by inhibiting sodium along with glucose transport in the proximal tubule, enhance the delivery of sodium to the juxtaglomerular apparatus, where increased sodium absorption by the macula densa cells leads to increased adenosine production, resulting in vasoconstriction of the afferent arteriole, decreased renal plasma flow, and reduced intraglomerular pressure. On a long-term basis, this process provides protection against DKD. This scenario is well established in both animal models of and humans with type 1 diabetes (28–31,86). However, a recent study suggests that, in human type 2 diabetes, the renal hemodynamic effects of SGLT2 inhibitors are caused by post-glomerular (efferent) vasodilation rather than pre-glomerular vasoconstriction (87).

Renal hypoxia is a characteristic feature of the diabetic kidney (31–37,88). Renal oxygen consumption is very high, second only to that of the heart. The hyperfiltering diabetic kidney, in combination with an elevated plasma glucose concentration, markedly increases the filtered load of glucose that is reabsorbed, along with sodium, primarily by the SGLT2 transporter in the proximal tubule by the sodium–potassium adenosine triphosphatase pump, and this requires energy in the form of ATP. Because the kidney has a limited ability to increase renal blood flow, oxygen demand exceeds supply, leading to hypoxia, which can promote DKD. The SGLT2 inhibitors, by blocking glucose–sodium cotransport in the proximal tubule, could possibly lead to a reduction in oxygen demand and prevention of DKD.

Podocytes are an integral component of the glomerular permselectivity barrier, and podocyte loss correlates strongly with albuminuria, which is toxic to the kidney, and with the decline in GFR in humans (89). Lipid accumulation in podocytes, as well as in mesangial and tubular cells, promotes glomerulosclerosis and tubulointerstitial renal disease (41–43). SGLT2 inhibitors cause a switch from glucose to fatty acid oxidation (64,67,68). By increasing fat oxidation in the kidney and reversing lipotoxicity, these drugs would be expected to improve podocyte function, decrease albuminuria, and slow the progression of DKD.

The end product of fatty acid oxidation is ketones, and, not surprisingly, SLGT2 inhibitors also increase the production of ketones (64,68,90). Ketones are freely taken up by the kidney in proportion to their plasma concentration. Furthermore, oxidation of ketones generates more ATP with less oxygen consumption than glucose. This process would reduce renal hypoxia, thereby contributing to the improvement in renal function. Because ketones are avidly taken up by the heart, provide a source of acetyl CoA for the tricarboxcillic acid cycle, and generate more ATP per molecule of oxygen consumed than glucose, a similar mechanism could explain the beneficial effects of the SGLT2 inhibitors on the heart (90).

Although diabetic ketoacidosis is rare in patients with type 2 diabetes who are treated with an SGLT2 inhibitor, it occurs in 3–4% of patients with type 1 diabetes who take an SGLT2 inhibitor (91). Although not approved for these patients in the United States, SGLT2 inhibitors are approved in some European countries for the treatment of patients with type 1 diabetes, and their use should be monitored closely in this group.

DKD accounts for ∼50% of patients entering ESRD programs. Hyperglycemia is the primary factor responsible for initiating the cascade of pathophysiologic events that result in ESRD. Therefore, tight glycemic control (A1C <6.5%) is essential to prevent the development of diabetic glomerulosclerosis and progressive decline in GFR. ACE inhibitors, ARBs, and MRAs all have been shown to slow the progression of DKD. Most recently, five large prospective studies (the EMPA-REG OUTCOME, CANVAS Program, DECLARE-TIMI 58, CREDENCE, and DAPA-CKD trials) have provided conclusive evidence that SGLT2 inhibitors, when added to ACE inhibitor/ARB therapy, reduce a renal composite outcome including progression to ESRD in patients with type 2 diabetes.

R.A.D. serves on advisory boards for AstraZeneca, Boehringer Ingelheim, Intarcia, Janssen, and Novo Nordisk; has received research support from AstraZeneca, Boehringer Ingelheim, Janssen, and Merck; and is a speaker’s bureau member for AstraZeneca and Novo Nordisk. No other potential conflicts of interest relevant to this article were reported.

Both authors researched data and wrote and edited the manuscript. R.A.D. is the guarantor of this work and, as such, had full access to all the data and takes responsibility for the integrity of this review.