Everyone has a threshold of renal tubular glucose concentration above which glucose appears in the urine (40). It was recognised that this threshold could be reduced with an apple tree bark extract called phlorizin (41, 42). Following the cloning of the SGLT 1 and SGLT 2 genes, phlorizin was characterised as a non-specific SGLT inhibitor. SGLT 1 is a low-capacity, high-affinity transporter located primarily in the gastrointestinal tract. It functions to absorb dietary glucose and is also expressed in the later renal proximal tubule segment (S3), where it is responsible for reabsorbing only ~3% of filtered glucose under normal physiological conditions. By contrast, SGLT 2 is a high-capacity, low-affinity transporter located mainly in the early renal proximal tubule segments. SGLT 2 is responsible for reabsorbing ~97% of filtered glucose. Inhibition of SGLT 2 therefore has a much larger effect on the glucose threshold than SGLT 1 inhibition. Dual SGLT 1/2 inhibitors have also been developed with the aim of increasing urinary glucose-lowering excretion because SGLT 1 has significant reserve capacity to reabsorb glucose if SGLT 2 is inhibited (43-45), but gut effects include a potential to cause diarrhea (34, 36).
For each SGLT 2 reabsorbed molecule of glucose, a sodium ion is co-transported. SGLT 2 inhibition therefore increases sodium delivery through the renal tubules to each nephron’s macula densa, and subsequently into the urine. The macula densa is the structural area of the early distal convoluted tubule which lies between the afferent and efferent arteriole of the mother glomerulus to that distal tubule. Sodium delivery to the macula densa results in changes to intraglomerular blood flow and pressure by means of changes to the calibre of the afferent arteriole, a homeostatic process referred to as tubuloglomerular feedback. High sodium delivery to the macula densa results in constriction of the afferent arteriole, which results in a decrease in glomerular blood flow and glomerular capillary pressure, whilst decreased sodium delivery has the opposite effect. Therefore, SGLT 2 inhibition augments macula densa tubuloglomerular feedback as well as generating a natriuresis to combine with the glycosuric osmotic diuresis. These sodium effects appear to be central to both the kidney and cardiovascular physiological effects of SGLT 2 inhibition.
1.3.1 SGLT 2 inhibitors’ glycaemic effects
Large FDA-mandated trials in type 2 DM populations were initiated in order to assess the cardiovascular safety of SGLT 2 inhibitors (47). These trials demonstrated that SGLT 2 inhibitors are non-inferior to placebo with respect to effects on major atherosclerotic (or adverse) cardiovascular events (MACE) with some trials also demonstrating superiority (26-28). Subsequently, the CREDENCE, DAPA-CKD and EMPA-KIDNEY trials have demonstrated SGLT 2 inhibition’s particular efficacy at reducing risk of kidney disease progression and heart failure hospitalisation in people with CKD down to an estimated glomerular filtration rate (eGFR) of 20 mL/min/1.73m2 (29, 35, 39, 48). The realisation from these trials that SGLT 2 inhibition confers cardiac and kidney protection independent of glycaemic effects and kidney function, with substantial benefits also evident in people without DM, has led to a shift in focus from purely lowering glycosylated haemoglobin (HbA1c) to disease risk modification. At the time of writing, there are thirteen reported large placebo-controlled trials in different settings, including trials in populations with type 2 DM, heart failure and CKD (see Table 1.1 for a listing, including key eligibility criteria and outcomes).
1.3.2 SGLT 2 inhibitors’ effects on kidney physiology
In people with CKD, the CREDENCE, DAPA-CKD and EMPA-KIDNEY trials have demonstrated SGLT 2 inhibition’s particular efficacy in reducing risk of kidney disease progression in people with type 2 DM and albuminuric diabetic kidney disease (29, 35, 48). Analyses of DAPA-CKD and EMPA-KIDNEY have shown that benefits extend to non-diabetic causes of CKD (39, 48, 49). Despite the attenuated ability of SGLT 2 inhibitors to lower glucose at reduced levels of kidney function, the relative benefits of SGLT 2 inhibition on kidney disease progression appear preserved in people without DM, and no threshold has been identified where relative benefits start to attenuate (29, 35, 39, 48, 49).
As described above, the key mechanism for renoprotection is considered to be through SGLT 2 inhibitor’s modulation of tubuloglomerular feedback, increased delivery of sodium to the macula densa and the enhancement of glomerular afferent arteriolar vasoconstriction (50, 51). The consequent reduction in kidney glomerular blood flow is believed to be responsible for the acute, reversible dip in kidney function, reductions in albuminuria, and slowing of kidney function decline following initiation of SGLT 2 inhibition (50). This modulation of kidney haemodynamics appears to persist in people with normoglycaemia (although it may be attenuated). Acute dips in eGFR on initiation of SGLT 2 inhibition are apparent in people without DM (52-54). Consequently, SGLT 2 inhibition is hypothesised to reduce intraglomerular hypertension and target the proposed final common pathway for kidney disease progression in people with or without DM (55).
SGLT 2 inhibition also reduces the risk of adverse events attributed to acute kidney injury (AKI) (49, 56), with a protective effect evident in the trials of people with type 2 DM (32, 56), heart failure (30, 34, 57) and CKD populations alike (29, 35, 36). The proposed protective mechanisms are reduced risk of ischaemic-reperfusion injury or renal tubular hypoxia from the lowered metabolic demand of inhibited co-transporters (58). Conceivably, a reduction in AKI risk may also translate into benefits on CKD progression, providing a mechanistic explanation for suggested beneficial effect of SGLT 2 inhibition on estimated glomerular filtration rate slopes in individuals with heart failure and without albuminuria (33, 57, 59, 60).
Establishing definitively whether or not albuminuria is a pre-requisite for kidney benefits of SGLT 2 inhibition was an important question to address as: (i) the majority of individuals with CKD do not have albuminuria, and (ii) if mechanistic theories about intraglomerular hypertension are correct, kidney benefits may be substantially different in the absence of albuminuria. The EMPA-KIDNEY trial, which included participants with a wide range of kidney function including people with CKD without albuminuria, demonstrated that level of albuminuria modified the effect of SGLT 2 inhibition on kidney disease progression, with large relative benefits in people with higher levels of albuminuria. However, the number of participants without albuminuria whose kidney disease progressed to at least a 40% decline in eGFR during the 2 years of follow-up was low, and exploratory analyses of eGFR slopes were required to assess the effects in this subgroup (39). These analyses demonstrated that the rate of kidney function decline was reduced by a meaningful degree in people both with and without albuminuria, and that longer treatment periods would be expected to lead to reductions in the risk of kidney failure.
Sections 2 and 3 of this guideline will provide a more detailed appraisal of trials assessing the effects of SGLT 2 inhibition on kidney disease, with section 7 also providing special consideration for people with type 1 DM, acute decompensated heart failure, or a functioning kidney transplant.
1.3.3 SGLT 2 inhibitors’ effects on cardiovascular physiology
The totality of the trial evidence shows that relative benefits of SGLT 2 inhibition on heart failure outcomes are both larger and more consistent than on MACE. Meta-analyses estimated that the risk of hospitalisation for heart failure was reduced by about one-third compared to placebo, and MACE risk is reduced by about 10% (61). This suggests that the natriuretic, osmotic diuretic and renoprotective effects of SGLT 2 inhibition are more effective at targeting heart failure pathophysiology than atherothrombotic risk. The effects of SGLT 2 inhibition on MACE risk may result from the more modest lowering effects of SGLT 2 inhibition on blood pressure, HbA1c and adiposity. For a more detailed review of the cardiac effects of SGLT 2 inhibition, see a 2021 update from the European Society of Cardiology ad-hoc task force (62).
1.3.4 SGLT-2 inhibitors’ effects on metabolism
SGLT 2 inhibitors have broad metabolic effects beyond lowering blood glucose. Glycosuria leads to increased plasma glucagon, which in turn increases hepatic glucose production in part by glycogenolysis (63, 64). Depletion of liver glycogen creates a fasting-like state initiating ketone generation from the liver as an alternative energy source (52). Randomised trials consistently show a dose-dependent reduction in weight (65). Whereas the early weight loss may be due to intra- and extra-vascular volume depletion (66), loss of adipose tissue does occur with longer-term treatment (67).
In addition to inducing a state of ketosis, SGLT 2 inhibition also reduces renal ammoniagenesis. This – in combination with ketosis – leads to urinary loss of bicarbonate which, combined with ketosis, may lower the threshold required to induce ketoacidosis in the presence of an additional insult (e.g. fasting or infection) (68). This accounts for both the increased risk of ketoacidosis in individuals taking SGLT 2 inhibitors and explains presentations of “euglycaemic ketoacidosis”. Ketoacidosis is a particular risk in individuals who have limited endogenous insulin production, and particularly those with type 1 DM (69). The benefit-risk ratio is therefore more finely balanced in people with type 1 DM. While use of certain SGLT 2 inhibitors has previously been granted by regulators at lower doses and under specialist supervision in this population (70, 71), authorisation for two SGLT 2 inhibitors have withdrawn from their indication for type 1 DM (72, 73). Section 7 of this guideline document introduces special consideration for people with type 1 DM.
The risk of severe hypoglycaemia caused by SGLT 2 inhibition is small (26-30, 32-39). Mechanistically, hypoglycaemia would not be expected because of the compensatory effects of intact SGLT 1 activity and hepatic gluconeogenesis (74). Hypoglycaemia on SGLT 2 inhibitors is therefore largely limited to individuals who are on concomitant hypoglycaemia-inducing medication (i.e. insulin or insulin secretagogues).
1.3.5 Potential adverse effects of SGLT 2 inhibitors
Mycotic genital infections are common in individuals with DM, but there is an increased risk associated with SGLT 2 inhibitor-induced glycosuria. The effect of SGLT 2 inhibition on these infections is large enough to have been apparent in the earlier smaller trials focusing on glycaemic control (75-77). Case reports of necrotizing fasciitis of the perineum (Fournier’s gangrene) attribute such devastating polymicrobial infections to SGLT 2 inhibitor use (78), but the rarity of the condition means there is insufficient randomised data to confirm or refute this hazard – a causal association remains unproven (28, 32, 35, 79). Large amounts of urinary glucose meant an increased incidence of urinary tract infection were expected and are listed in the labels for all SGLT 2 inhibitors. However, this effect is small and only apparent when results from all the large randomised trials (26-30, 32-39) are combined in meta-analysis (49, 80).
The CANVAS program, which tested canagliflozin in individuals with type 2 DM at high risk of cardiovascular disease, raised two new safety considerations: an excess of bone fractures and separately, an increased risk of lower limb amputation were identified (27). Post hoc biological rationales have been proposed for these effects of canagliflozin (81), but a chance finding is a plausible alternative explanation. Excess of lower limb amputation or bone fracture has not been observed in the other large trials of SGLT 2 inhibition, including the CREDENCE trial, which assessed canagliflozin in individuals with type 2 DM and CKD (29, 49).
Section 5 of this guideline document will provide a more detailed appraisal of the trials assessing the effects of SGLT 2 inhibition on metabolic and safety outcomes, with section 6 providing supporting information for people with CKD being offered SGLT 2 inhibition.