A Review of the Efficacy and Safety of Sodium–Glucose Cotransporter 2 Inhibitors: A Focus on Diabetic Ketoacidosis

Sodium–glucose cotransporter 1 (SGLT1) is predominantly located in the small intestine, but is also expressed in the kidneys, trachea, heart, and colon (4,5). In the kidneys, SGLT1 is primarily located in the S3 segment of the proximal convoluted tubule (PCT) (4). SGLT2 is expressed in the kidneys and primarily located in the S1 and S2 segments of the PCT (4,5). In normoglycemic adults, about 180 g of glucose (Figure 1) is filtered per day in the glomerulus, and most is reabsorbed (4,6). In people with diabetes, reabsorption of glucose is increased compared to people without diabetes (7,8). SGLT1 and SGLT2 are located in the apical membrane and facilitate the transport of glucose with sodium from the renal tubular lumen into the cells (Figure 2) (4).

Phlorizin (Figure 1), an O-glucose derivative/O-glycoside, was discovered in 1835 from apple tree bark (9,10). Phlorizin is a dual SGLT inhibitor, inhibiting both SGLT1 and SGLT2 (10). Canagliflozin, dapagliflozin, and empagliflozin are C-glucose derivatives and selectively inhibit SGLT2 (11). SGLT2 inhibitors are structurally similar to glucose, as shown in Figure 1, and thereby competitively inhibit glucose, leading to increased levels of glucose in the urine (5,6). Clinically available SGLT2 inhibitors block ∼30–50% of filtered glucose (6).

The efficacy of SGLT2 inhibitors as monotherapy, dual and triple oral therapy, and in combination with insulin has been established in randomized, controlled trials. A1C reduction associated with SGLT2 inhibitors ranges from 0.5–1% and varies based on dose, severity of diabetes, and other patient-specific factors (12–15).

Because of the insulin-independent mechanism of action of SGLT2 inhibitors, they may be used at all stages of type 2 diabetes, including more severe stages, in which endogenous insulin secretion has declined significantly. This mechanism explains why the risk of hypoglycemia is rare, although it may still occur when an SGLT2 inhibitor is used in combination with an insulin secretagogue or exogenous insulin (16). SGLT2 inhibitors are also associated with a consistent reduction of systolic and diastolic blood pressure by 2–4 and 1–2 mmHg, respectively, as a result of their osmotic diuretic effect (12,17). Weight loss of ∼2 kg has been observed with SGLT2 inhibitors as a result of their glucosuric effect, and even greater weight reductions have been observed in patients with a higher baseline BMI (14,18).

As with many other antidiabetic agents, data on microvascular outcomes with SGLT2 inhibitors are lacking. However, macrovascular and mortality outcomes with empagliflozin are now available, and cardiovascular and mortality trials of canagliflozin and dapagliflozin are underway. EMPA-REG OUTCOME, a 3-year trial in patients with type 2 diabetes and high cardiovascular risk, found that empagliflozin significantly decreased the primary composite outcome with a number needed to treat of 63, driven by a reduction in cardiovascular death (19). Neither myocardial infarction, stroke, nor hospitalization for unstable angina was reduced compared to placebo. Hospitalization for heart failure was 2.7% with empagliflozin compared to 4.1% with placebo (P = 0.002) (19). Although the precise explanation for empagliflozin’s beneficial clinical outcomes is unknown, it is likely multifactorial. Potential reasons include the agent’s effects on arterial stiffness, cardiac function, and cardiorenal function (19,20). Empagliflozin’s ability to reduce albuminuria, uric acid, body weight, visceral adipose tissue, and blood pressure may provide additional mechanisms (19,21,22). CANVAS is an ongoing randomized, double-blind, placebo-controlled trial studying the effect of canagliflozin on cardiovascular outcomes and death in patients with uncontrolled type 2 diabetes and a history of cardiovascular events (23). DECLARE-TIMI 58 is an ongoing randomized, double-blind, placebo-controlled trial investigating the effect of dapagliflozin on cardiovascular death, myocardial infarction, and stroke in patients ≥40 years of age with type 2 diabetes (24). These trials will provide more insight regarding the cardiovascular effects of SGLT2 inhibitors.

SGLT2 inhibitors are generally well tolerated, but some disadvantages are associated with this therapy. An increase in urogenital infections has been observed because of their effect on increased urinary glucose. A pooled analysis of clinical trials found 11 and 4% increased risks of genital mycotic infection in women and men, respectively, compared to placebo. Events were generally mild to moderate in severity and responded to standard therapy (25). The FDA has since issued a warning regarding the risk of urinary tract infections leading to urosepsis and pyelonephritis with SGLT2 inhibitors (26). Health care providers should ask whether patients have a history of urogenital infections before initiating SGLT2 inhibitor therapy.

SGLT2 inhibitors are also associated with a small, reversible decrease in estimated glomerular filtration rate (eGFR), thereby decreasing the magnitude of their effect on glucose excretion and thus their efficacy as renal function declines (21,22,27). Hence, canagliflozin, dapagliflozin, and empagliflozin have variable dosing adjustments and restrictions based on eGFR. The FDA strengthened a warning on the labels of canagliflozin and dapagliflozin in June 2016 after receiving 101 case reports of acute kidney injury and recommends considering predisposing factors before initiating these therapies (28). However, this warning does not apply to empagliflozin, which recently was reported in a subanalysis of EMPA-REG OUTCOME to be associated with a slower progression of kidney disease compared to placebo in patients with mild renal dysfunction (29). It is unknown whether this is a class effect. The concept of renal protection relates to SGLT2 inhibitors’ ability to decrease uric acid levels, tubular glucose toxicity, and diabetes-induced hyperfiltration (30). The CREDENCE trial, now underway, will shed light on whether canagliflozin has beneficial renal effects in patients with type 2 diabetes and stage 2 or 3 chronic kidney disease (31).

Because of SGLT2 inhibitors’ effects on blood pressure, their use may lead to postural hypotension and dizziness, particularly in elderly patients, those taking loop diuretics, or those with tenuous intravascular volume. Therefore, caution and dose adjustments may be warranted in such patients (32,33). Pooled trial data from long-term canagliflozin therapy showed an increase in bone fracture rates, leading the FDA to issue a new warning in September 2015 for decreased bone mineral density and to strengthen its warning about increased bone fracture risk (34). SGLT2 inhibitors increase serum phosphate levels, likely via tubular reabsorption, thereby increasing both parathyroid hormone (PTH) and fibroblast growth factor (FGF) 23. PTH and FGF 23 promote phosphaturia and have opposite effects on vitamin D metabolism, although evidence has shown that SGLT2 inhibitors decrease mean 1,25 dihydroxyvitamin D levels (35). Neither dapagliflozin nor empagliflozin carry bone fracture risk warnings (36,37).

There have also been concerns of bladder and breast cancer with dapagliflozin because it was associated with a nonsignificant increase in phase 2 and 3 trials. However, this may be attributable to detection bias (38–40). Molecular and animal evidence does not suggest a positive link between SGLT2 inhibitor exposure and cancer risk (41). Minimal increases in LDL cholesterol also have been noted with SGLT2 inhibitors, potentially resulting from metabolic changes such as increased lipoprotein lipase activity, but the exact mechanism is unknown (22,42).

Diabetic ketoacidosis (DKA) is a potentially life-threatening complication in people with diabetes, predominantly those with type 1 diabetes. DKA typically is defined as the triad of hyperglycemia (blood glucose >250 mg/dL), anion gap metabolic acidosis, and the presence of urine or plasma ketones (43). Euglycemic DKA (euDKA) is rare and defined as DKA with a blood glucose level ≤250 mg/dL. It may be precipitated by incomplete DKA treatment or reduced insulin dose, food restriction, alcohol consumption, or inhibition of gluconeogenesis (44). Metabolic changes during pregnancy may also predispose patients to euDKA (45).

After reviewing FDA Adverse Event Reporting System database entries since the approval of canagliflozin in March 2013, the FDA issued a warning in May 2015 about the risk of DKA associated with SGLT2 inhibitors. The report found 73 cases of DKA in patients with type 1 diabetes or type 2 diabetes treated with SGLT2 inhibitors, specifically 44 cases in type 2 diabetes, 16 in type 1 diabetes, 13 unspecified, and 1 in a patient with latent autoimmune diabetes in adults (LADA) (26). Canagliflozin, dapagliflozin, and empagliflozin were associated with 21, 4, and 4 DKA cases, respectively (26). Concomitant dehydration, infection, and changes in insulin dose were reported in 73% of the cases. Management took place in emergency departments or inpatient settings in all of the cases, and the FDA identified possible risk factors, including infection, low-carbohydrate diet or reduced caloric intake, alcohol use, and reduced dose or discontinuation of insulin or oral insulin secretagogue therapy (26). The FDA has added this warning to the labels of all SGLT2 inhibitors, and post-marketing pharmacovigilance studies are ongoing.

Although the exact mechanism is not fully understood, the following proposed mechanisms may explain how SGLT2 inhibitors cause euDKA (46):

1. Reduced insulin levels and enhanced glucagon secretion. SGLT2 inhibitors act by blocking the reabsorption of filtered glucose in the PCT, leading to an increased excretion of glucose in the urine and decreased levels of glucose in the blood (47,48). The lower blood glucose levels result in the reduced secretion of insulin from the pancreatic β-cells.This enhances the secretion of glucagon from the pancreatic α-cells, which is referred to as an indirect effect of SGLT2 inhibitors on glucagon (49,50). Additionally, in vitro human and in vivo mice studies show that dapagliflozin directly acts on the pancreatic α-cells and triggers the secretion of glucagon, providing evidence that SGLT2 inhibitors are indeed α-cell secretagogues (51). Eventually, reduced insulin and increased glucagon levels will initiate lipolysis in adipose tissues and β-oxidation of fatty acids, leading to the formation of ketone bodies in the liver and potentially causing euDKA (47,49). Patients with reduced insulin levels, whether from a reduction in insulin dose or an insulin deficiency, may be at increased risk.

2. Reduced renal clearance of ketone bodies. SGLT2 inhibitors also may contribute to decreased excretion of ketone bodies synthesized in the body because phlorizin decreases the renal clearance of ketone bodies (46).

In people with type 2 diabetes, no clear signal to suggest DKA was noted in large clinical development programs for any of the three marketed SGLT2 inhibitors. Several case reports are now noted in the literature (Table 1) (52–56). As evidenced by case reports, sudden withdrawal of insulin therapy or secretagogues during the initiation of SGLT2 inhibitors may increase the risk of DKA. Additionally, patients following a low-carbohydrate diet may be at risk, and ensuring appropriate hydration is essential because dehydration may lead to acceleration of ketogenesis (54). Two cases of euDKA developed in the postoperative period, and further research will be needed to provide recommendations for SGLT2 inhibitor therapy in the pre- and postoperative periods (56).

Further information is available from trial data. Erondu et al. (57) reviewed DKA events in the canagliflozin type 2 diabetes clinical program. Twelve of the 17,596 patients developed DKA or related events while receiving canagliflozin. However, half of the 12 patients were reported to have type 1 diabetes or LADA. The EMPA-REG OUTCOME study conducted for 3 years with 7,020 patients found no difference in the rate of DKA with empagliflozin compared to placebo (19). Based on >18,000 patients with type 2 diabetes in randomized, controlled study programs, the frequency of DKA in those exposed to dapagliflozin is <0.1% (44).

Although SGLT2 inhibitors are not FDA-approved for use in people with type 1 diabetes, they have been used off label in practice. An 18-week phase 2 study of 351 patients with type 1 diabetes assessed the efficacy and safety of canagliflozin as add-on treatment to insulin (58). Twelve patients in the canagliflozin group developed DKA that required hospitalization. No patients in the placebo group experienced a ketone-related adverse effect. Five of the 12 patients had a blood glucose level <250 mg/dL at the time of hospitalization (58). Additional cases of canagliflozin-associated euDKA in type 1 diabetes have been reported in the literature (56). Precipitating factors such as illness, change in diet, increased activity, and reduction or omission of insulin led to the DKA episodes (56,58).

SGLT2 inhibitors are a second-line or later therapy option for the management of type 2 diabetes. However, a recent FDA warning regarding SGLT2-associated DKA has raised concerns. The proposed mechanism relates to SGLT2 inhibitors’ indirect effects on reducing endogenous insulin levels and enhancing glucagon secretion, while also reducing renal ketone clearance. Although case reports have demonstrated an increased risk of DKA in both type 1 and type 2 diabetes, larger randomized, controlled trials are expected to provide greater understanding. Patients should be educated about the risks of DKA. Interruption of SGLT2 inhibitor treatment may be warranted during periods of prolonged fasting due to illness or surgery, low-carbohydrate diet, dehydration, stress, or changes in insulin or insulin secretagogue medications.

No potential conflicts of interest relevant to this article were reported.