#NephJC Chat
Tuesday October 6 2020 9 pm Eastern
Wednesday October 7 2020 9 pm IST
Wednesday October 7 9 2020 pm BST
N Engl J Med. September 24, 2020 DOI: 10.1056/NEJMoa2024816
Dapagliflozin in Patients with Chronic Kidney Disease.
Hiddo J.L. Heerspink, Ph.D., Bergur V. Stefánsson, M.D., Ricardo Correa-Rotter, M.D., Glenn M. Chertow, M.D., Tom Greene, Ph.D., Fan-Fan Hou, M.D., Johannes F.E. Mann, M.D., John J.V. McMurray, M.D., Magnus Lindberg, M.Sc., Peter Rossing, M.D., C. David Sjöström, M.D., Roberto D. Toto, M.D., Anna-Maria Langkilde, M.D., and David C. Wheeler, M.D., for the DAPA-CKD Trial Committees and Investigators*
PMID: 32970396
See also: DAPA-CKD Protocol paper; DAPA-CKD Baseline data
Introduction
Yawn. One more sodium glucose co-transporter 2 inhibitor (SGLT2i) trial showing benefit. What else is new? In the past, we discussed the EMPA-REG results (Wanner et al, NEJM 2016, NephJC summary) which were confirmed with CANVAS (Neal et al NEJM 2017, NephJC summary). DECLARE-TIMI 58 (Wiviott et al, NEJM 2019) showed similar though slightly less impressive results. But these were cardiovascular trials, primarily designed to satisfy the FDA guidance on proving cardiovascular safety, which serendipitously revealed not just safety, but reduced cardiovascular and kidney outcomes with these drugs. CREDENCE (Perkovic et al, NEJM 2019, NephJC summary) was the first study to focus on patients at high risk of kidney outcomes and promote those outcomes to the primary end-point. After CREDENCE, SGLT2i were firmly entrenched as the standard of care in the treatment of diabetic kidney disease. A few months ago, we even discussed some fascinating mechanistic data (Griffin et al, Circulation 2020, NephJC summary) on how these drugs even work. Aren’t we done yet?
Having conquered end organ damage in diabetes, the unstoppable SGLT2i juggernaut is now heading for….non-diabetic kidney (and heart) disease.
Nope. Having conquered end organ damage in diabetes, the unstoppable SGLT2i juggernaut is now heading for….non-diabetic kidney (and heart) disease. Earlier this year, NephMadness (see scouting report) crowned SGLT2i without Diabetes as the champion. Then DAPA-HF trial (McMurray et al, NEJM 2019) and the EMPEROR-HF trial (Packer et al, NEJM 2020) drew our attention to the benefits of SGLT2i in halting heart failure progression and lowering cardiovascular (CV) mortality in patients with diabetes mellitus as well as without diabetes mellitus.
How can that be? SGLT2 inhibitors lower blood glucose levels modestly by inhibiting glucose reabsorption in the proximal tubule leading to urinary glucose and sodium excretion. When we analyze the impact on kidney function, similar to RAAS blockers, the commencement of treatment with SGLT2 inhibitors is characterized by acute, dose-dependent reductions in eGFR ( up to ~5 mL·min–1·1.73 m2 ) over several weeks (Heerspink et al, Circulation 2016). After this initial reduction, the eGFR subsequently tends to return toward the baseline and remains stable over time. But here, the initial decline in eGFR is due to afferent arteriolar vasoconstriction through a tubuloglomerular feedback mechanism. SGLT2i restore delivery of sodium to the macula densa, promoting adenosine production and increased afferent arteriolar tone. Thus this dip in the eGFR reflects favorable hemodynamic changes in the glomerulus. This reduction in hyperfiltration and intraglomerular pressure would also suppress downstream glomerular fibrosis and inflammation and the favorable hemodynamic effect is the plausible explanation for their benefit in diabetic CKD. Could it also preserve kidney function in patients with kidney diseases due to causes other than diabetes mellitus? And how do you then explain its benefit in heart failure - seen in non-diabetic patients as well? There’s no tubuloglomerular feedback involved in the Frank-Starling curve! If it's merely the natriuretic effect at play - why don’t we see those kinds of benefits with other natriuretics?
They do lower blood pressure (and don’t get started with whether BP reduction is beneficial for the kidneys) by about 4/2 mm Hg (Baker et al, J Am Heart Assoc, 2017). If RAS blockade works in non-diabetic CKD as well, particularly when there is evidence of intraglomerular hypertension and albuminuria, then SGLTi might work via a similar mechanism in these settings. Blockade of glucose and sodium reabsorption in the PCT can also lead to other potential beneficial processes like improvement in mitochondrial functions, decrease hypoxic damage to proximal tubular cells resulting in less intrarenal inflammation.
Animal data in non-diabetic model is mixed, see below from Dekkers et al (NDT 2020).
Table 2 from Dekkers et al, NDT 2020
In animals which are not rodents, a 6-week trial with dapagliflozin reported no effect on proteinuria in patients with CKD without diabetes, but did report an acute and reversible decline in mGFR and a reduction in body weight (DIAMOND, Cherney et al, Lancet Diab Endo 2020). But the real test is in longer term clinical trials, examining clinical outcomes.
This leads us to Dapagliflozin and Prevention of Adverse outcomes in CKD (DAPA-CKD) trial, to explore whether the promising beneficial effects of SGLT2 inhibitors extend to non-diabetic kidney disease patients.
The Study
Methods
This was a randomized, double-blind, placebo-controlled, multicenter clinical trial conducted at 386 sites in 21 countries. (NCT03036150)
Here is the visual abstract from NDT just for the design:
Study Population
Inclusion criteria
Age: more than 18 years
eGFR: ≥25 and ≤75 mL/min/1.73m2 (CKD-EPI Formula)
UACR between ≥200 mg/g and ≤5000 mg/g
Stable patient on maximum tolerated daily dose of angiotensin-converting enzyme (ACE) inhibitor or angiotensin receptor blocker (ARB) for at least 4 weeks, if not medically contraindicated
Exclusion criteria
Type 1 DM
CKD due to ADPKD or ARPKD, lupus nephritis or ANCA-associated vasculitis
Receiving cytotoxic therapy, immunosuppressive therapy, or other immunotherapy within 6 months prior to enrollment
History of organ transplantation
Receiving therapy with SGLT2 inhibitor within 8 weeks prior to enrollment
NYHA class IV CHF at the time of enrolment MI, unstable angina, stroke, or TIA within 12 weeks prior to enrollment
PCI or CABG or valvular repair/replacement within 12 weeks prior to enrollment
Active malignancy requiring treatment ( except successfully treated basal cell or treated squamous cell carcinoma)
ALT/AST - >3x the ULN, or total bilirubin >2x ULN
Pregnancy or breastfeeding
Intervention
Participants were randomly assigned 1:1 to dapagliflozin 10 mg/day or matched placebo. Randomization was stratified according to the diagnosis of type 2 DM (yes or no) and UACR (≤1000 or >1000). After randomization, in-person trial visits were performed at 2 weeks, at 2, 4, and 8 months, and at 4-month intervals thereafter.
Randomization was monitored to ensure that a minimum of 30% of the participants were recruited to either the population with type 2 DM or the population without diabetes.
Figure 1 from Heerspink et al, NDT 2020
Trial endpoints
Primary endpoint:
This was a composite endpoint, of the time to first occurrence of any of the following:
≥50% eGFR decline from baseline OR
Onset of ESKD defined as eGFR <15 mL/min/1.73 m2, need for chronic dialysis and renal transplantation OR
Death from renal or cardiovascular causes
Secondary endpoints:
Time to a composite renal endpoint
≥50% eGFR decline from baseline
ESRD defined as eGFR <15 mL/min/1.73 m2, need for chronic dialysis or renal transplantation
Renal death (which was defined as death from ESKD, but where dialysis was deliberately withheld: eg refused, considered futile or not available)
Time to the first occurrence of either cardiovascular death or hospitalization for heart failure
Time to death from any cause
Statistical analysis
681 primary outcome events were needed to detect a 22% lower relative risk in the dapagliflozin group than in the placebo group (hazard ratio of 0.78), with 90% power using a two-sided alpha level of 0.05 and assuming an annual event rate for the primary outcome of 7.5% in the placebo group.
A Cox proportional-hazards regression model that was stratified according to type 2 DM and UACR and adjusted for the baseline eGFR was used to estimate (hazard ratio) HR and 95% CI.
Funding
The trial was funded by AstraZeneca, the manufacturer of Dapagliflozin. An executive committee consisting of nine academic members and two non-voting employees of the sponsor was responsible for the design and oversight of the trial and the reporting of the results. The analyses that were conducted by the sponsor were replicated by an independent academic group at the University Medical Center Groningen. The first draft of the manuscript was written by the first and last authors and was revised by the coauthors. The decision to submit the manuscript for publication was made jointly by all the authors. Four of the fourteen authors are employees of the sponsor.
Results
Overall, 4304 participants were enrolled into the trial. The ‘figure 1’ is sadly 😥 relegated to the supplementary appendix as figure S2, seen below:
Figure S2 from DAPA-CKD, NEJM 2020
Table 1 depicts the baseline demographic characteristics of patients.
Table 1 from DAPA-CKD, NEJM 2020
The mean (±SD) age was 61.8±12.1 years. The mean eGFR was 43.1±12.4 ml /min/1.73 m2, the median UACR was 949, and of more interest to us, 2906 (67.5%) patients had type 2 Diabetes Mellitus. More on that is shown in the baseline data paper (Wheeler et al, NDT 2020).
Figure 2 from Wheeler et al, NDT 2020
Characteristics of study participants: (A) study participants by diabetes status, (B) investigator-reported causes of CKD in all study participants and (C) investigator-reported causes of CKD in study participants with T2D
Primary outcomes
The plan was to continue the trial until 681 outcomes occurred. However, the trial was halted early, on the basis of the DSMB recommendation, when only 408 of the 681 expected outcomes had occurred, in view of the overwhelming efficacy. The primary composite outcome occurred in 197 participants (9.2%) in the dapagliflozin group and 312 participants (14.5%) in the placebo group (HR, 0.61; 95% [CI], 0.51 to 0.72; P<0.001) - so 39% RR reduction compared to the expected 22% reduction.
Figure 1A from DAPA-CKD, NEJM 2020
Secondary outcome
The HR for the kidney composite outcome was 0.56 (95% CI, 0.45 to 0.68; P<0.001)
Figure 1B from DAPA-CKD, NEJM 2020
The components of the outcomes can be seen in the table below.
Table 2 from DAPA CKD, NEJM 2020
The HR for the composite of death from CV causes or hospitalization for heart failure was 0.71 (95% CI, 0.55 to 0.92; P=0.009). There were 101 deaths (4.7%) from any cause in the dapagliflozin group and 146 (6.8%) in the placebo group (HR, 0.69; 95% CI, 0.53 to 0.88; P=0.004). Very impressive - and if you calculate, there were less non-CV and non-renal deaths in the dapaglifozin group compared to placebo as well.
Figure 1 C and D from DAPA-CKD, NEJM 2020
Subgroup analysis
The effect of dapagliflozin on the primary outcome was generally consistent across prespecified subgroups. In participants with type 2 diabetes mellitus.
Figure 2 from DAPA-CKD, NEJM 2020
Change from baseline in eGFR
The least-squares mean (±SE) eGFR slopes from baseline to 30 months in the dapagliflozin and placebo groups were –2.86±0.11 and –3.79±0.11 ml/min/ 1.73 m2 / year, respectively, resulting in a between-group difference of 0.93 ml per minute per 1.73 m2 per year (95% CI, 0.61 to 1.25) (Fig. 3). The annual change in the mean eGFR was smaller with dapagliflozin than with placebo (–1.67±0.11 and –3.59±0.11 ml /min/1.73 m2, respectively), for a between-group difference of 1.92 ml per minute per 1.73 m2 per year (95% CI, 1.61 to 2.24).
Figure 3 from DAPA-CKD, NEJM 2020
Safety Overall, the trial demonstrates a beneficial safety profile of dapagliflozin. In particular, there were no cases of diabetic ketoacidosis, and hypoglycemic episodes did not occur in participants without diabetes.
Table 2 section showing safety data from DAPA-CKD, NEJM 2020
Discussion
Are these results valid?
Early Stoppage: The trial was stopped early - which is a completely reasonable thing to do in this scenario - what is the point of continuing the trial when it is clear the results are clearly going to be positive? Better close out and announce the results. The only drawback is that this may exaggerate the magnitude of the benefit (See Bassler et al, JAMA 2010). The smaller number of events also means less events and less certainty in the robustness of the subgroup results.
Non Cardiovascular death reduction: The reduction in CV death is understandable, and so is renal death. However, the reduction in non-cardiovascular, non-renal deaths (~ 30 less, RR 0.57) is impressive but puzzling. Is that a chance finding, exaggerated from the early stoppage? Knowing the exact causes of deaths would be useful here. Or it just might be that DAPA-CKD and EMPAREG are outliers - see below via Christos Argyropoulos
Figure from Twitter metananalysis, Argyropoulos, 2020
What makes the DAPA-CKD trial different from other SGLT2 inhibitor trials?
To date, CREDENCE was the only SGLT2 inhibitor trial to recruit patients with both type 2 DM and CKD, with the lower eGFR cutoff for inclusion was 30 ml /min/1.73 m2. The enrollment of a broader spectrum of patients (CKD, with and without diabetes) was an important characteristic of the DAPA CKD study. 32.5 percent of patients with CKD did not have diabetes mellitus and 14.5% had an eGFR below 30 ml per minute per 1.73 m2. A different clinically meaningful component i.e sustained eGFR of <15 mL/min/1.73 m2 was also included in the definition of ESRD.
Figure modified from (Heerspink et al. NDT 2020)
More forthcoming SGLT2 inhibitors trials in CKD
DAPA-CKD solidifies the benefit of SGLTi in diabetic kidney disease. How strong are the effects in non-DM CKD? Here,, the results of another forthcoming dedicated CKD outcome trial EMPA-KIDNEY trial are eagerly awaited. The EMPA-KIDNEY trial extends the inclusion criteria further and enrolls patients ( with or without DM) with eGFR: ≥20-45 or eGFR ≥45 to <90 with UACR ≥200 mg/g, with an emphasis on the lower end of GFR and albuminuria.
Further Questions
Drug effect or class effect?
There had been some concern after DECLARE-TIMI that the effect of dapaglifozin was not as impressive as empaglifozin and canaglifozin. Those concerns are now washed away. See a nice twitter meta-analysis for all the outcomes laid out in this thread below. Though the results from VERTIS-CV (Cannon et al, NEJM 2020) for ertuglifozin are somewhat underwhelming.
Do these results apply to all non-diabetic CKD or just proteinuric CKD?
DAPA-CKD enrolled patients with proteinuria. Non-proteinuric CKD, notably polycystic kidney disease is excluded. The sheer progression of kidney function is very slow in these - so it is harder to show a difference. There may be other considerations in causing natriuresis along with aquaresis in PKD. Inflammatory glomerulonephritis were also excluded: Lupus and vasculitis. No reason to think SGLT2i have any role here, yet.
How low can we go for GFR with respect to initiating SGLT2i?
Published data from clinical trials report empaglifozin initiation down to GFR 30 (EMPAREG) ; canaglifozin down to GFR 30 (CREDENCE) and now dapaglifozin down to 25. Drug labels aside, there is nothing from the pharmacology to suggest the GFR threshold for starting these drugs is not a class effect. So many might feel comfortable starting these drugs at GFRs > 25. EMPA-KIDNEY might bring that threshold down to GFR 20. Below that, whether one should start these drugs might also be similar to the issue of starting RAS blockade in CKD 4 or 5. Is it too late? It might be too late for kidney disease - but will starting these drugs at low GFR still help with heart failure and CV benefit?
Both in CREDENCE and here in DAPACKD, the drugs were continued all the way down to dialysis - and hence this practice definitely has support. There is no need to stop at a GFR threshold once they are started.
Side effects: what's the final word on amputations, euglycemic DKA and infections?
CANVAS raised the ugly spectre of amputations - which seems to have not been substantiated in the other trials. The risk of genital infections seems to be real more than UTIs. Euglycemic DKAs as well were more common (with a low absolute rate) in CREDENCE, but not seen here. Something to keep in mind with adverse effects is that patients in clinical trials are followed closely with excellent adherence to protocols - so in the non-trial clinical setting, adverse events are usually more common.
Conclusion
Regardless of the presence or absence of diabetes, dapagliflozin significantly lowered the renal or cardiovascular events and deaths when compared to placebo. DAPA-CKD is the first SGLT2 inhibitor trial to declare a favorable clinical outcome in non-diabetic CKD patients. Let us hope that the advent of the SGLT2 inhibitor bestows a new era in the field of nephrology.
Summary prepared by
Priti Meena
Department of Nephrology,
Sir Gangaram hospital New Delhi, India
NSMC Intern, Class of 2020