Sodium-glucose cotransporter 2 inhibitors for diabetic kidney disease: a primer for deprescribing

Abstract

Chronic kidney disease (CKD) is a critical global public health problem associated with high morbidity and mortality, poorer quality of life and increased health care expenditures. CKD and its associated comorbidities are one of the most complex clinical constellations to manage. Treatments for CKD and its comorbidities lead to polypharmacy, which exponentiates the morbidity and mortality. Sodium-glucose cotransporter 2 inhibitors (SGLT2is) have shown remarkable benefits in cardiovascular and renal protection in patients with type 2 diabetes mellitus (T2DM). The pleiotropic effects of SGLT2is beyond glycosuria suggest a promising role in reducing polypharmacy in diabetic CKD, but the potential adverse effects of SGLT2is should also be considered. In this review, we present a typical case of a patient with multiple comorbidities seen in a CKD clinic, highlighting the polypharmacy and complexity in the management of proteinuria, hyperkalemia, volume overload, hyperuricemia, hypoglycemia and obesity. We review the cardiovascular and renal protection effects of SGLT2is in the context of clinical trials and current guidelines. We then discuss the roles of SGLT2is in the management of associated comorbidities and review the adverse effects and controversies of SGLT2is. We conclude with a proposal for deprescribing principles when initiating SGLT2is in patients with diabetic CKD.

CASE PRESENTATION

A 67-year-old morbidly obese male with a history of hypertension, type 2 diabetes mellitus (T2DM), systolic heart failure and hyperuricemia was followed in the renal clinic for chronic kidney disease (CKD) Stage 3a with nephrotic-range proteinuria. He had a recent kidney biopsy for increasing proteinuria and serum creatinine, which revealed diabetic nephropathy with chronic active interstitial nephritis. He was previously taken off a blocker of the renin–angiotensin–aldosterone system (RAAS) because of multiple episodes of hyperkalemia. Despite being on a low potassium diet, furosemide and patiromer, his potassium remained >5 mEq/L, which precluded reintroduction of the RAAS blockade medications. Also, his endocrinologist had recently intensified his diabetic regimen with insulin due to poor hemoglobin A1c control, and since then he has experienced weight gain and more frequent episodes of hypoglycemia. He also required up-titration of furosemide due to fluid retention in the lower extremities. The higher dose of furosemide precipitated a gout attack in his right knee, for which he took a course of steroids and was started on allopurinol. He reported persistent right knee pain from the gout attack, which severely limited his mobility and ability to exercise and left him feeling overwhelmed by his growing medication list. In the renal clinic, he expressed frustration that his renal function had declined further despite all his efforts to adhere to the medical advice of his multiple health care providers.

INTRODUCTION

CKD is a critical global public health problem associated with high morbidity and mortality, poorer quality of life and increased health care expenditures [1]. Comorbid conditions like diabetes, hypertension, hyperlipidemia, hyperuricemia, heart failure and cardiovascular disease are highly prevalent in CKD [2, 3] and associated with increased mortality [4, 5]. This constellation of conditions can be difficult to manage, often leading to polypharmacy in an attempt to manage comorbidities and mitigate the progression of CKD [6–8]. Indeed, as kidney function declines, patients experience additional complications, including anemia, bone mineral disorders, acidosis, hypervolemia and cardiovascular complications, all of which often require medication therapy. Most CKD patients take an average of 8–13 medications [9]. The number of prescribed medications is a recognized predictor of prescribing problems, including inappropriate dosage, drug–drug interactions and drug–disease interactions [10]. The use of multiple complex medication regimens in CKD increases drug-related problems [11] and inappropriate drug use is associated with 40% higher mortality in patients with CKD compared with those with preserved kidney function [12].

Our index case highlights a prescribing cascade, a process whereby the side effects of drugs result in more prescriptions, which causes additional side effects and unanticipated drug interactions [13]. Prescribing cascades similar to the example above are not uncommon in managing diabetic kidney disease.

Balancing the management of CKD, including associated comorbidities and complications, with the minimization of necessary and appropriate medications is challenging. However, the nephrologist now has sodium-glucose cotransporter 2 inhibitors (SGLT2is), a novel class of diabetic medications with many potentially helpful uses. Large clinical trials of SGLT2is have demonstrated remarkable benefit among patients with T2DM in reducing the risk of cardiovascular death, heart failure hospitalization and progression of renal disease [14]. The pleiotropic effects of SGLT2is beyond glycosuria suggest a promising role in managing multiple problems with a single once-daily pill, yet the efficacy and safety profile in moderate CKD is less clear. In this review we present a typical case of a patient with multiple comorbidities seen in CKD clinic, highlighting the complexity in management and resultant polypharmacy. We discuss the current evidence and guidelines for the use of SGLT2is in patients with diabetic CKD. We review the roles that SGLT2is may play in mitigating CKD complications, managing comorbidities and decreasing medication burden in this population, as well as the potential adverse effects of SGLT2is. We conclude with a proposal for safer deprescribing methods when initiating SGLT2is in the renal clinic.

Cardiovascular and renal protective effects of SGLT2is and current guideline recommendations

The US Food and Drug Administration (FDA) approved the use of SGLT2is for T2DM and cardiovascular risk reduction in patients with an estimated glomerular filtration rate (eGFR) >45mL/min/1.73 m². The characteristics of the study populations and the cardiovascular and renal outcomes of the pivotal trials of SGLT2is [15–18] are summarized in Table 1. The study population typically represented a group of obese patients with long-standing, poorly controlled T2DM and with various degrees of preexisting cardiovascular and renal disease who were receiving the standard of care for the management of diabetes and cardiovascular disease. Notably, the Empagliflozin Cardiovascular Outcome Event Trial in Type 2 Diabetes Mellitus Patients–Removing Excess Glucose (EMPA-REG) and Canagliflozin Cardiovascular Assessment Study (CANVAS) study populations had high cardiovascular disease burden, the Evaluation of the Effects of Canagliflozin on Renal and Cardiovascular Outcomes in Participants With Diabetic Nephropathy (CREDENCE) trial population had high renal disease burden and the Multicenter Trial to Evaluate the Effect of Dapagliflozin on the Incidence of Cardiovascular Events (DECLARE-TIMI58) had the lowest cardiovascular and renal disease burdens. The EMPA-REG, CANVAS and CREDENCE trials demonstrated reduced major adverse cardiovascular events (MACEs), whereas the DECLARE-TIMI58 did not, probably reflecting less severe cardiovascular burden in the study population. Unequivocally, all four trials showed reduced heart failure hospitalizations irrespective of cardiovascular disease burden. All three SGLT2is reduced the renal composite outcomes compared with placebo. The CREDENCE trial observed the highest renal event rates and absolute risk reduction, followed by the EMPA-REG, CANVAS and DECLARE-TIMI58 trials, which is a downtrend likely corresponding to the degree of renal disease burden. In summary, SGLT2is are effective in the prevention of heart failure hospitalizations and progression of CKD and reducing cardiovascular events in patients with high cardiovascular risk.

Table 1.

Summary of key study population characteristics and outcomes in the pivotal clinical trials of SGLT2is

Characteristics	EMPA-REG	CANVAS	DECLARE-TIMI58	CREDENCE
Study drug	Empagliflozin	Canagliflozin	Dapagliflozin	Canagliflozin
Sample size	7020	10 142	17 160	4401
eGFR inclusion criteria (mL/min/1.73 m2)	>30	>30	>60	30–90
Albuminuria inclusion criteria (mg/g)	NA	NA	NA	300–5000
Average age (years)	63	63	64	63
Duration of diabetes (years)	14	14	11	16
Percent with prior cardiovascular disease	100	66	40	51
Percent with preexisting heart failure	10	14	10	15
Body mass index (kg/m2)	31	32	32	31
Blood pressure (mmHg)	135/77	137/78	135/85	140/78
Glycated hemoglobin (%)	8.1	8.2	8.3	8.3
Baseline eGFR (mL/min/1.73 m2)	74	76	85	56
Percent with eGFR <60 mL/min/1.73 m2	26	21	7	41
Percent with albuminuria	40	31	NA	99
Percent on metformin	74	77	82	58
Percent on insulin	48	50	41	66
Percent on sulfonylurea	43	43	43	29
Percent on ACEi or ARB	81	80	81	100
Percent on diuretics	44	44	41	47
Percent on statin	77	75	75	69
Outcomes (placebo versus experiment, rate/1000 patient-year)
Progression of proteinuriaa	118 versus 104*	129 versus 89*	NA	NA
Renal composite outcomesb	11.5 versus 6.3*	9.0 versus 5.5*	7.0 versus 3.7*	40.4 versus 27.0*
Three-point MACEc	43.9 versus 37.4*	31.5 versus 26.9*	24.2 versus 22.6	48.7 versus 38.7*
Hospitalization for Heart failure	14.5 versus 9.4*	8.7 versus 5.5*	8.5 versus 6.2*	25.3 versus 15.7*

^*P < 0.05.

^aIncident albuminuria or a 30% increase in albuminuria or progression to macroalbuminuria.

^bDoubling of serum creatinine level accompanied by eGFR ≤45 mL/min/1.73 m², initiation of renal replacement therapy or death from renal disease (EMPA-REG); 40% reduction in eGFR, renal replacement therapy or renal death (CANVAS); ≥40% decrease in eGFR to <60 mL/min/1.73 m², ESRD or death from renal cause (DECLARE-TIMI58).

^cDeath from cardiovascular causes, nonfatal myocardial infarction or nonfatal stroke.

The 2019 American Diabetes Association (ADA) guidelines recommend SGLT2is as the preferred class of agents to be added after metformin in patients with T2DM and a history of heart failure or CKD, if eGFR is >30 mL/min/1.73 m² [19, 20]. SGLT2is are also recommended as one of the two preferred agents (the other is GLP-1 receptor agonists) to be added after metformin in patients with atherosclerotic cardiovascular disease (ASCVD). Furthermore, clinicians should consider switching from other diabetes agents without cardiovascular or renal benefit to SGLT2is, even in patients who are currently at their target hemoglobin A1c level. In addition, the ADA recommends prescribing SGLT2is in patients who would benefit clinically from weight loss and to minimize the risk of hypoglycemia. The ADA also acknowledges that in those without established cardiovascular disease or CKD, SGLT2is are the last recommended class of agents for use because of the high drug price. The cost-effectiveness of SGLT2is compared with other oral hypoglycemic drugs in patients with cardiovascular disease, heart failure or CKD is yet to be determined.

The pleiotropic effects of SGLT2is on polypharmacy in diabetic CKD

The highly prevalent comorbidities and associated polypharmacy highlight the complexity of care for patients with diabetic CKD (Table 2) [21]. The pleiotropic effects of SGLT2is show great promise in addressing these complex polypharmacy issues in diabetic CKD, including the management of proteinuria, hyperkalemia, volume overload, hyperuricemia, hypoglycemia, obesity and anemia (Table 3).

Table 2.

The prescribing cascade in the management of CKD

Medications	Used for	Unintended consequence	Added medication burden
ACEi/ARB/spironolactone	Proteinuria, hypertension, heart failure	Hyperkalemia	Patiromer, diuretics
Loop and thiazide diuretics	Hypertension, heart failure, hyperkalemia	Hyperuricemia	Urate-lowering agents
		Hyperglycemia	Diabetic agents
Insulin/insulinotropic agents	Diabetes	Weight gain	Weight reduction medications
		Sodium retention	Diuretics, antihypertensives
		Hypoglycemia	Not applicable

Table 3.

The pleiotropic effects of SGLT2is

Clinical indication	Current therapy	Possible benefits when combining or substituting with SGLT2is besides cardiac and renal protection	Optimization Strategy
Prevent adverse renal and cardiovascular events	ACEi/ARB, spironolactone [22, 23]	Synergy in proteinuria reduction	Coprescribe
		Reduce hyperkalemia from RAAS blockers
Secondary hyperparathyroidism	Active vitamin D analogs [24]	Mitigate the SGLT2i-mediated up-regulation of FGF23 and suppression of 1,25-dihydroxyvitamin D [25]	Coprescribe if indicated
Volume overload	Loop and thiazide diuretics	Synergy in natriuresis	Coprescribe if indicated with dose adjustment
		Less metabolic derangement
Glucose control	GLP-1 RAs	Additional glycemic control with low risk of hypoglycemia	Coprescribe if indicated
	DPP-4 inhibitors	Additional glycemic control with low risk of hypoglycemia	Coprescribe if indicated
	Thiazolidinediones	Avoid sodium retention and exacerbation of heart failure	Deprescribe
	Insulin/insulinotropic agents	Weight loss, less sodium retention and lower hypoglycemic risk	Deprescribe
Prevent hyperkalemia	Loop and thiazide diuretics	Less metabolic derangement	Deprescribe
	Patiromer	Lower medication cost and simpler dosing schedule	Deprescribe
Hypertension	CCB, α1-blocker, clonidine	No	Deprescribe

BP, blood pressure; CCB, calcium channel blocker; ESA, EPO-stimulating agent.

SGLT2is as an add-on or rescue therapy to RAAS blockers for proteinuria

The angiotensin-converting enzyme inhibitors (ACEis) and angiotensin II receptor blockers (ARBs) are the cornerstone classes of medications to reduce proteinuria and slow CKD progression. However, the rate of hyperkalemia from RAAS blockade is 3-fold higher in patients with CKD than in those without CKD [26]. Incident hyperkalemia is associated with a dose reduction or discontinuation of RAAS blockers [27], increased emergency room visits [27] and higher mortality [26]. The trend of ACEi/ARB usage rates in patients with diabetes and CKD has decreased since the early 2000s [28, 29], possibly due to resultant hyperkalemia. Dietary consultation on low potassium diets and coprescribing of kaliuretic diuretics and intestinal potassium binders are the current potassium-lowering strategies to allow reintroduction of ACEis/ARBs.

The efficacy of proteinuria reduction of SGLT2is is ∼30–40% in proteinuric diabetic kidney disease with the eGFR range from 30 to 90 mL/min/1.73 m² [30, 31], which is comparable to ACEi/ARB or spironolactone monotherapy [22, 32]. A 30% reduction in proteinuria is associated with a reduced incidence of CKD and end-stage renal disease (ESRD) [33, 34]. In the SGLT2i clinical trials, 80–100% of patients were on an ACEi or ARB agent and the addition of an SGLT2i further reduced the rate of incident or worsening nephropathy by 30–45% [15–18], suggesting that SGLT2is should be used as an add-on agent to treat diabetic kidney disease. In patients who are intolerant of RAAS blockade, SGLT2is represent alternative renoprotective agents with less incident hyperkalemia in the clinical trials [15, 35–37], even in patients with moderate renal insufficiency [15, 18, 38]. SGLT2i-induced natriuresis in the proximal tubule [39] and the induction of aldosterone [40, 41] promote kaliuresis from the distal tubule, but a clinically relevant potassium-lowering effect was not demonstrated in clinical trials [36].

Besides the potassium-neutral proteinuria reduction effect, SGLT2is raised the hematocrit by 2–3% [40, 42] in both CKD and non-CKD patients [35]. This observation is attributed not only to its hemoconcentration effect, but also erythropoiesis [40] due to increased endogenous erythropoietin (EPO) production [43], which may be related to mitigation of renal cortex hypoxia and improvement of pericyte function [44].

SGLT2i as an add-on to overcome diuretic resistance

Diuretic resistance is commonly seen in CKD patients and leads to poor outcomes. Poor drug absorption due to intestinal edema, reduced effective drug concentration due to proteinuria and impaired tubular secretion and compensatory sodium reabsorption from unblocked sodium transporters (such as Na⁺/H⁺ exchanger 3, Na-Cl cotransporter and SGLT2) all contribute to diuretic resistance [45]. Many strategies to overcome diuretic resistance have been attempted, but none have demonstrated long-term clinical benefit [46]. Loop diuretics are the most commonly used drugs for decongestion in CKD. Chronic loop diuretic use increases distal nephron sodium retention via the Na-Cl cotransporter. The combination of thiazide diuretics and loop diuretics has been shown to promote natriuresis and decongestion in loop diuretic–resistant patients, but long-term benefit in reducing cardiovascular mortality remains uncertain and metabolic and electrolytes disorders are common [47].

Multiple large clinical trials have demonstrated that SGLT2is significantly reduce heart failure hospitalizations regardless of a history of ASCVD or existing heart failure [14]. A recent study demonstrated a mutually adaptive synergy between SGLT2is and loop diuretics, and combination therapy was superior to either drug alone in promoting natriuresis [39]. Interestingly, despite diminished blood glucose–lowering effects of SGLT2is as renal function worsens, the antihypertensive and cardioprotective effects persisted across all eGFR groups [48, 49], suggesting that SGLT2is maintain natriuresis in patients with reduced eGFR despite lower drug delivery to the proximal tubules. Studies are currently under way examining SGLT2is [50] or thiazide diuretics [51] in combination with loop diuretics on heart failure outcomes and mortality. In euvolemic patients who are on a stable diuretic regimen, the addition of SGLT2i may require down-titration or discontinuation of existing diuretics to avoid overdiuresis.

SGLT2is mitigate hyperuricemia and hyperglycemia from loop and thiazide diuretics

Loop and thiazide diuretics are commonly used for salt-sensitive hypertension, volume management or hyperkalemia. Loop and thiazide diuretics cause increased uric acid net absorption from the proximal tubules by direct effects on urate transporters [52, 53] and by the indirect effect on volume contraction. Compared with nondiuretic antihypertensives, the use of diuretics increases the serum uric acid level by 0.7–0.9 mg/dL [54, 55] and increases the risk of gout attacks [55], which can potentially lead to worse cardiovascular and renal outcomes [56]. Also, loop and thiazide diuretics are associated with insulin resistance [57] and dose-dependent increased risk of incident diabetes [58, 59], probably related to compensatory sodium reabsorption coupled with increased glucose reabsorption via SGLT2 [39]. Since hyperuricemia, metabolic syndrome and diabetes are closely related and coexist in patients with CKD, the frequent use of loop or thiazide diuretics for blood pressure and volume management places CKD patients at risk for developing new comorbidities.

SGLT2i-induced glycosuria improves hyperglycemia and modulates the urate transporter GLU9 to promote uric acid excretion [60] and reduce serum uric acid level by ∼0.8 mg/dL [61]. Substituting thiazide diuretics with an SGLT2i can improve metabolic derangements without affecting blood pressure control [62]. Therefore the addition of an SGLT2i could mitigate the metabolic derangements of loop and thiazide diuretics and allow for deprescribing of uric acid–lowering agents.

SGLT2is as diabetic agents with low hypoglycemic risk and weight reduction effect

Insulin and sulfonylureas are well-known to cause hypoglycemia [19, 63–66], especially in patients with CKD [67], whereas the newer classes of glucose-lowering agents, such as SGLT2i, glucagon-like peptide-1 receptor agonists (GLP-1 RA) and dipeptidyl peptidase 4 inhibitors pose a lower risk of hypoglycemia but are underprescribed in CKD patients [68]. Diabetes and CKD together compound the risk for incident hypoglycemia, which is associated with excess mortality [69].

The higher incidence of hypoglycemia in CKD patients is likely related to reduced insulin clearance and renal gluconeogenesis as renal function declines [70]. Moreover, insulin directly suppresses gluconeogenesis in the proximal tubules, while inhibition of SGLT2 enhances gluconeogenesis via activation of sirtuin1/peroxisome proliferator-activated receptor γ coactivator-1α [71], thus providing mechanistic insights into the higher risk of hypoglycemia seen with insulin and sulfonylureas, especially in patients with CKD, and justification for substituting them with SGLT2is. SGLT2is have been used in patients with eGFR >30 mL/min/1.73 m² in clinical trials [15, 17, 18] and have shown a low risk of hypoglycemia [72]. As an add-on to insulin-based or sulfonylurea-based glucose-lowering regimens, SGLT2is improved glycemic control with a lower risk of hypoglycemia [73, 74].

Weight reduction improves cardiovascular and renal risk factors [75, 76] and independently improves renal outcomes [77]. Insulin and sulfonylureas cause weight gain, which often leads to serial escalations in the treatment of diabetes, joint pain, hypertension, hyperlipidemia or proteinuria. SGLT2is can reduce weight by 2–4 kg on average [72] due to total body water reduction upon initiation and then by increasing fat metabolism with chronic use [78]. The weight loss effect is maintained in mild to moderate CKD [49] and attenuated in moderate to severe CKD [79].

The adverse effects of SGLT2is

While SGLT2is have demonstrated great promise for cardiovascular and renal protection, the potential adverse effects of SGLT2is should not be overlooked. The common adverse effects include osmotic diuresis, volume depletion and genital fungal infections. Less common adverse effects that have been reported in some clinical trials but not seen in meta-analyses [38, 72] include urinary tract infection, diabetic ketoacidosis, lower extremities amputation and bone fracture. Diabetic ketoacidosis was only seen in the CREDENCE and DECLARE-TIMI58 trials and the lower extremities amputation was only seen in the CANVAS trial, but not in the CREDENCE, EMPA-REG or DECLARE-TIMI58 trials. Last but not least, 55 cases of necrotizing fasciitis of the perineum were reported in case series [80] but not seen in clinical trials.

Of note, the risk of bone fracture is controversial. In both healthy volunteers [25] and T2DM patients with moderate CKD [81], canagliflozin increased serum phosphate, which stimulated fibroblast growth factor 23 (FGF23) and decreased 1,25-hydroxyvitamin D [25] and accelerated the loss of hip bone mineral density [82]. A higher risk of bone fracture was seen in the CANVAS trial. However, the risk of bone fracture was neither higher in the SGLT2i arm in the CREDENCE, EMPA-REG or DECLARE-TIMI58 trials [15, 16, 18] nor in the meta-analysis of the safety profile of SGLT2is in CKD patients [38]. The magnitude and variability of bone mineral adverse effects of SGLT2is are yet to be determined in the postmarket safety surveillance. It is advisable to monitor patient’s parathyroid hormone (PTH) level while on SGLT2is and treat with active vitamin D analogs if the PTH level uptrends above the goal.

In summary, starting SGLT2is in a highly select group of CKD patients after meticulously weighing the risks and benefits and frequent monitoring of volume status, genital and foot hygiene, electrolytes, kidney function and bone mineral biomarkers are the keys to ensuring the safety of patients on SGLT2is.

The controversy around using SGLT2is for renoprotection in moderate to severe diabetic CKD

The glucose-lowering effect of SGLT2is when the eGFR is <45 mL/min/1.73 m² is almost negligible. At present, the FDA does not recommend the use of canagliflozin, empagliflozin and dapagliflozin when the eGFR is <45 mL/min/1.73 m². However, diabetic patients with moderate to severe proteinuric CKD have a very high risk of progression to ESRD despite the current standard of care, which is a substantial unmet medical need. The CREDENCE trial and the subgroup analysis of the EMPA-REG and CANVAS trials showed that the cardiovascular risk reduction, heart failure hospitalization reduction and composite renal outcome reduction were maintained in patients with an eGFR of 30–45 mL/min/1.73 m² and the safety profile was similar to the patients with preserved renal function, which argues for the ‘off-label’ use of SGLT2is in T2DM patients with proteinuria and an eGFR of 30–45 mL/min/1.73 m². The most updated ADA guidelines in June 2019 recommend the use of SGLT2is in diabetic CKD patients with an eGFR >30 mL/min/1.73 m² [20]. Other clinical trials of SGLT2is that target cardiovascular and renal outcomes in both diabetic and nondiabetic CKD patients are currently under way [EMPA-Kidney (The Study of Heart and Kidney Protection With Empagliflozin; NCT03594110) and DAPA-CKD (A Study to Evaluate the Effect of Dapagliflozin on Renal Outcomes and Cardiovascular Mortality in Patients With Chronic Kidney Disease; NCT03036150)].

Deprescribing principles when using SGLT2is in patients with diabetic CKD

The pleiotropic effects of SGLT2is can potentially break the prescribing cascade that contributes to polypharmacy (Figure 1). The following proposed deprescribing principles for SGLT2i use are intended only for diabetic kidney disease.

FIGURE 1.

Prescribing cascade in CKD and SGLT2is for deprescribing. A hypothetical prescribing cascade after prescribing ACEi/ARB for proteinuria in a diabetic CKD3a patient. The multifaceted effects of SGLT2is show great promise in the management of diabetic CKD and its comorbidities. ESAs, EPO-stimulating agents.

The first principle is to ‘think SGLT2i first’. Adding an SGLT2i or switching to an SGLT2i is recommended in type 2 diabetic patients with an eGFR >30 mL/min/1.73 m² and with established ASCVD, heart failure or CKD [19, 20]. For CKD patients with adequate glycemic control who are seen in a CKD clinic, initiation of or switching to an SGLT2i by the treating nephrologist may reduce the likelihood of polypharmacy and hypoglycemia, providing more individualized diabetes and renal care.

The second principle is to ‘coprescribe other agents for synergy or mitigation of side effects’ (Table 3). For example, coprescribing RAAS blockers has synergy in proteinuria reduction and renal protection and mitigates the hyperkalemia caused by RAAS blockers; coprescribing active vitamin D analogs mitigates the induction of FGF23 and suppression of 1,25-dihydroxyvitamin D by SGLT2is [25].

The third principle is to ‘deprescribe less desirable agents when possible’. The less desirable agents may have a similar treatment effect but less desirable features, such as diabetic agents with no cardiovascular or renal protection benefit, dyssynergy with existing treatment, safety concerns or agents that may yield poor adherence due to regimen complexity or side effects (Table 3). From our experience, empagliflozin initiation necessitated adjustments to diabetes, hypertension and diuretic regimens in almost all patients. Whether SGLT2i use reduces the medication burden warrants further investigation. The deprescribing process requires careful planning, active surveillance and interdisciplinary communications [83].

The final principle is that ‘when taking a holistic approach to medication management, safety comes first’. This principle requires the establishment of individualized treatment goals and consideration of the whole patient at the time of prescription or deprescription [67, 84]. Monitor closely for adverse effects when starting or stopping medications. We provide a list of practical considerations (Table 4) in the decision-making process of medication management in patients with CKD, specifically related to the use of SGLT2is.

Table 4.

Practical considerations for when to defer initiation or withhold SGLT2is

Clinical conditions	Concerns	Recommended actions
Cognitive impairment	Unable to follow instructions for surveillance and medication titration; at risk for dehydration	Avoid SGLT2is; assess cognitive impairment
Poorly controlled hyperglycemia and a history of diabetic ketoacidosis	Poor adherence; risk of hyperglycemia-induced dehydration; risk of ketoacidosis	Avoid SGLT2is; referral to diabetologist
Unhealed diabetic foot wound	Risk of amputation	Avoid SGLT2is; referral to podiatry
History of frequent urinary tract infections, indwelling urinary catheter or self-catheterization	Risk of urinary tract infection	Avoid SGLT2is
Frequent vaginal candidal infections or unable to maintain genital hygiene for medical (e.g. phimosis) or social (e.g. insecure housing) reasons	Risk of genital fungal infection	Avoid SGLT2is
Peripheral vascular disease	Risk of amputation	Use with caution; frequent foot exam while on SGLT2is; referral to vascular surgery
Osteoporosis or renal osteodystrophy	Risk of bone fracture	Use with caution; monitor PTH and total vitamin D levels; referral to osteoporosis specialist for fracture risk assessment
Dynamic volume status (poor oral intake, diarrhea) or labile blood pressure	Dehydration and hypotension	Use with caution; deprescribe other diuretics and antihypertensives
Frequent nocturia	Patient intolerant of polyuria	Reduce dose or withhold

The outcome of the index clinical case

After the introduction of empagliflozin to the patient’s medication regimen, he experienced significant diuresis in the first 4 weeks and his serum creatinine increased from 1.21 (eGFR = 62 mL/min/1.73 m²) to 1.58 mg/dL (eGFR = 45 mL/min/1.73 m²) and potassium fell <4 mEq/L. We discontinued his patiromer and reduced the furosemide dose. His serum creatinine returned to baseline and his potassium stabilized at 4.0–4.8 mEq/L. His proteinuria decreased significantly from 7.8 to 3.8 g and we successfully reintroduced a low-dose ARB. His uric acid level was 2.1 mg/dL at the 6-month follow-up visit and allopurinol was discontinued. Over 6 months, he lost 11 kg and his insulin dose was reduced, with hemoglobin A1c remaining at the goal. In response to an increase in his PTH level, we started him on low-dose calcitriol. In summary, after the introduction of empagliflozin, we discontinued or reduced four medications, reintroduced one renoprotective medication and initiated one medication for bone health. The patient tolerated the SGLT2i well and was thankful for the reduction in his proteinuria as well as the number of medications.

CONCLUSION

Diabetic CKD is complex to manage because of the high prevalence of comorbidities and resultant polypharmacy. The pleiotropic effects of SGLT2is show remarkable cardiovascular and renal benefits and great promise to address polypharmacy issues in diabetic CKD. Meticulous patient selection, individualized treatment plans and careful surveillance for adverse effects are the keys to ensure safety in patients on SGLT2is. SGLT2is are recommended in appropriate patients even when glycemic control is adequate. Further study on whether SGLT2is decrease medication burden for patients with diabetic kidney disease is warranted.

FUNDING

J.L. is supported by the Ben J. Lipps Research Fellowship from the American Society of Nephrology. J.M.P. is supported by National Institutes of Health NIDDK K23 grant (1K23DK100447).

CONFLICT OF INTEREST STATEMENT

None declared. The results presented in this article have not been published previously in whole or part, except in abstract format.