In early trials of management strategies in type 2 diabetes, intensive glycaemic control was found to reduce incident microvascular complications, with less convincing benefit on macrovascular complications including cardiovascular mortality. Furthermore, the cardiovascular profile of some oral hypoglycaemic agents was somewhat uncertain. In this context, a meta-analysis on the peroxisome proliferator activated receptor gamma agonist rosiglitazone was published showing that across 42 trials, cardiovascular mortality was increased 64% in the treatment arm - prompting an advisory label on the product and apprehension of its use in patients with preexisting cardiovascular background.
Subsequently, the Food and Drug Administration (FDA) issued recommendations in 2008, that approval of new antidiabetic drugs required the demonstration of satisfactory cardiovascular safety, generally via a primary endpoint comprising a composite of cardiovascular mortality, myocardial infarction, and stroke. Large scale randomised cardiovascular outcome trials (CVOTS) followed of dipeptidyl peptidase 4 (DPP4) inhibitors, glucagon linked peptide-1 agonists, and SGLT2i.
Although these CVOTs focused on cardiovascular endpoints, a range of secondary outcomes examined effect on kidney outcomes. The first of these was the Empagliflozin Cardiovascular Outcome Event Trial in Type 2 diabetes Mellitus Patients (EMPA-REG Outcome) trial, where empagliflozin reduced incipient nephropathy and worsening kidney function. Then in the CANagliflozin cardiovascular Assessment Study (CANVAS) trial, canagliflozin reduced progression of albuminuria, slowed the reduction in eGFR and need for kidney failure requiring replacement therapy by 40%. Finally, in the Dapagliflozin Effect on Cardiovascular Events–Thrombolysis in Myocardial Infarction 58 (DECLARE–TIMI 58) trial, dapagliflozin reduced progression of a renal composite outcome by 24%.
Renal Outcome Trials
The Canagliflozin and Renal Events in Diabetes with Established Nephropathy Clinical Evaluation (CREDENCE) trial of canagliflozin 100 mg daily versus placebo in 4401 patients with diabetes was the first the published CVOT to predeclare a renal primary endpoint: a composite of doubling of serum creatinine, end stage kidney disease, or death from renal or cardiovascular causes. The trial participants had mean estimated glomerular filtration rate (eGFR) 56 ml/min and required proteinuria for inclusion (uACR 300–5000 mg/g). The trial was terminated early after 2.4 years, at which point canagliflozin reduced the primary end point by 30%. The Dapagliflozin and Prevention of Adverse Outcomes in Chronic Kidney Disease (DAPA-CKD) trial evaluated the effect of dapagliflozin 10 mg daily versus placebo on a similar combined renal end point, and for the first time, recruited patients without diabetes, who comprised a third of the study population. A total of 4304 patients were included, with mean eGFR 43 ml/min and again, inclusion required significant proteinuria (uACR 200– 5000 mg/g). Dapagliflozin reduced the primary end point by 39%, with similar benefit seen in patients both with and without diabetes. These trials are the first studies to demonstrate a reduction in mortality with an intervention in patients with CKD. There is no evidence as yet for benefit in patients with CKD without proteinuria at this stage, but EMPA-KIDNEY (NCT03594110) is including these patients.
A recent meta-analysis has concluded that for every 1000 patients with CKD without diabetes treated with SGLT2i for a year, 19 first kidney disease events (decline in eGFR >40%) were avoided.
These clinical trial outcomes have been borne out in real world settings and therefore it is the responsibility of the renal community to interpret and implement the findings of these landmark trials and overcome any barriers, perceived or actual.
Proposed Mechanisms of Action of Sodium-glucose Cotransporter 2 Inhibitors
In a reversal of the typical drug development timeline, translational research is playing catch up to explain the overwhelming clinical benefit of SGLT2i in CVOTs, and indeed much of the pleiotropic effects remain poorly explained.
SGLT2 is a glucose transporter that is localised in the proximal convoluted tubule of the kidney and resorbs the majority of glucose filtered at the glomerulus. SGLT2 are upregulated in diabetes which can result in a maladaptive response that exacerbates hyperglycaemia.
The SGLT2i act in the proximal tubule to inhibit sodium and glucose reabsorption via SGLT2, increasing distal delivery of sodium, chloride and water to the macula densa. This results in tubuloglomerular feedback with glomerular afferent arteriolar vasodilatation, reduction in intraglomerular perfusion pressure and reduction in filtration. The osmotic effects of increased sodium and glucose delivery continue in the distal nephron, despite likely upregulation of distal sodium reabsorption, and an osmotic diuresis is seen. Familial renal glycosuria is an inherited absence of SGLT2. In this condition, there can rarely be resultant hypotension or hypoglycemia, but the condition is otherwise relatively benign, with no association with CKD.
Several clinical effects of these drugs are predictable. On average, weight loss of around 2 kg is expected. The initial volume loss appears to then be maintained by loss of abdominal girth and an increase in lipolysis, maintaining the initial weight loss. Glycaemic control is improved in diabetes, with an average reduction in HbA1c of around 0.6%, although this benefit is lost with reductions in eGFR. Systolic blood pressure is mildly reduced by around 3mmHg.[22,23]
Reduction in proteinuria is seen with SGLT2i, presumably as a consequence of the reduction in intraglomerular perfusion pressure. An average reduction of 30% in uACR is seen by 26 weeks. Beneficial effects on kidney outcomes in clinical trials are independent of reduction in proteinuria. What is known now is the greater the proportional reduction in proteinuria, the greater the reduction in risk of cardiovascular or renal end points.
Some clinical effects are unexpected and as yet unexplained, such as the reduction in nonosmotically active sodium stores. The effect on interstitial versus intravascular sodium is one of ways in which SGLT2i differ from conventional diuretics.
Implementation of the Evidence
The early focus of trials was on diabetic kidney disease,[13,10] but subgroup analysis from DAPA-CKD suggests that the next opportunity for nephrology is using SGLT2i in the setting of nondiabetic proteinuric CKD.[27,28] The following seven suggestions will hopefully allow the impressive results of these large randomised controlled trials to be implemented with improved outcomes for patients (Figure 1). We additionally highlight some ongoing controversies and how these agents might be used in glomerular disease.
Seven steps to successful SGLT2i use. ACEi, angiotensin converting enzyme inhibitor; APKD, adult polycystic kidney disease; ARB, angiotensin receptor blocker; eGFR, estimated glomerular filtration rate; GN, glomerulonephritis; K, potassium; KFRT, kidney failure requiring replacement therapy; SGLT2i, sodium-glucose cotransporter 2 inhibitors.
Should these medicines be used as monotherapy? At present, there is no evidence for use of SGLT2i in patients with CKD as monotherapy. In CREDENCE and DAPA-CKD patients were required to be prescribed maximally tolerated angiotensin converting enzyme inhibitor (ACEi)/angiotensin receptor blocker (ARB) for inclusion. In both these trials >98% of participants were treated with these agents at baseline. Therefore, we would not recommend instituting SGLT2i without ACEi/ARB for renoprotection.
Should I check eGFR and potassium shortly after starting an SGLT2i? The effects of these medicines on glomerular haemodynamics mean that an initial reduction in eGFR of 20–30% is to be expected, much like can be seen with ACEi. In fact, data have shown that an early drop in eGFR is associated with better outcomes in patients prescribed SGLT2i; therefore, checking for a drop in renal function is not necessary but could be considered for individual prognostication. No effect on potassium was seen in trials with these medicines, despite patients being prescribed ACEi/ARB.
Will all patients with primary glomerulonephritis benefit from SGLT2i? In DAPD-CKD, patients with primary glomerulonephritis were included. 6.3%  had a diagnosis of immunoglobulin A nephropathy (IgAN). This means that more patients were trialled with SGLT2i in DAPA-CKD than with methylprednisolone in TESTING (n = 262) or immunosuppression in STOP-IgA (n = 162), large multicentre IgAN trials. Other primary glomerulonephritis including focal segmental glomerulosclerosis were also included in DAPA-CKD. Patients receiving active immunosuppression were not included and similarly, patients with lupus nephritis and anti-neutrophil cytoplasmic antibody (ANCA)-associated vasculitis were excluded. Therefore, there are some important exclusion criteria at present for these drugs. Patients with CKD without proteinuria were also not studied but EMPA-KIDNEY will hopefully provide further information on this cohort.
Where do these drugs fit in the spectrum of therapy for patients with glomerular diseases where immunosuppression is contemplated? For patients with an immune-mediated primary glomerulonephritis and a healthy cardiometabolic phenotype, such as a young person with primary focal segmental glomerulosclerosis, we believe that management should continue as before with prompt and aggressive immunosuppression with the primary aim of prompt remission induction in keeping with existing Kidney Disease Improving Global Outcomes (KDIGO) guidelines. For patients in whom prompt remission is less likely, for example, a patient with primary membranous nephropathy who does not fit the criteria for immunosuppression but is likely to remain proteinuric for some time, then the role for SGLT2i in this setting should be established (Figure 2). At present, for patients with IgA nephropathy, management is largely around blood pressure control and maximising RAS inhibition, with steroid therapy considered for a minority with heavier proteinuria despite these measures. Given the results of DAPA-CKD, we believe any clinical trial in IgA nephropathy should include SGLT2i and maximised ACEi/ARB as standard of care.
Do I need to reduce the patient's diuretic dose when I start these medicines? Initial intravascular volume reduction is seen with SGLT2i, although longer term the effects are seen less on the intravascular volume and more on the interstitium. Data from Dapagliflozin and Prevention of Adverse Outcomes in Heart Failure (DAPA-HF) suggest diuretic doses may not need to be adjusted but in a real-world setting, close observation is recommended and individualised management decisions instituted.
Is there a level of eGFR below which the benefits are not seen? The licence for level of eGFR to start these medicines varies between countries, but in the clinical trials, these medicines were continued until the patient reached end stage kidney disease, assuming they were tolerated and benefits were seen regardless of level of eGFR at initiation into the trial.[30,35]
What are the sick day rules for these medicines? The risk of euglycaemic ketoacidosis with these medicines is well described, although no instances have been observed in trials of patients without diabetes. However, it is recommended to advise patients to discontinue their medicines if they are fasting or unable to eat, undergoing a major operation or following a low carbohydrate diet. They should recommence once they are reliably eating. There is also a three-fold increased risk of fungal genital infections with SGLT2i.
Role of SGLT2i in management of glomerular diseases. DMN, diabetes mellitus nephropathy; FSGS, focal segmental glomerulosclerosis; RASi, renin angiotensin system inhibition; SGLT2i, sodium-glucose cotransporter 2 inhibitors.
There are likely to be further trials of these medications in different clinical situations. It would be helpful to know if the benefits of these medicines are additive or synergistic to RAS inhibition and whether SGLT2i as standalone treatment is of benefit. Secondly, whether they can be used in the setting of acute volume overload, as an additional transporter not currently targeted by conventional diuretics. We note pilot trials have been performed in patients with acute decompensated heart failure and suggest these agents are likely to be safe and efficacious in this setting, although larger trials are required. Interestingly, this class of drugs has been associated with a reduced incidence of acute kidney injury, another as yet unexplained finding which may lead to an expansion in indication, for example, in prevention of acute kidney injury in high-risk settings. Furthermore, it remains uncertain at what stage of CKD SGLT2i should be started in patients with or without diabetes, with absolute benefit likely to be less with well preserved renal function. It also remains uncertain whether patients should follow a reduced sodium diet to gain further benefit.
What is clear is that any future studies in glomerular disease will require to incorporate SGLT2i as existing standard of care. There is now a clear basis that in emerging trials of novel immunotherapies for glomerular disease, patients should be well established on SGLT2i in addition to RAS inhibition, before being randomised.
Curr Opin Nephrol Hypertens. 2022;31(3):272-277. © 2022 Lippincott Williams & Wilkins