#NephJC Chat
Tuesday November 5th, 2024 at 9 pm EST
Wednesday November 6th, 2024 at 9 pm IST
N Engl J Med 2024 Oct 25:10.1056/NEJMoa2409183.
doi: 10.1056/NEJMoa2409183. Online ahead of print.
Long-Term Effects of Empagliflozin in Patients with Chronic Kidney Disease
EMPA-KIDNEY Collaborative Group; William G Herrington, Natalie Staplin, Nikita Agrawal, Christoph Wanner, Jennifer B Green, Sibylle J Hauske, Jonathan R Emberson, David Preiss, Parminder Judge, Doreen Zhu, Rejive Dayanandan, Ryoki Arimoto, Kaitlin J Mayne, Sarah Y A Ng, Emily Sammons, Michael Hill, Will Stevens, Karl Wallendszus, Susanne Brenner, Alfred K Cheung, Zhi-Hong Liu, Jing Li, Lai Seong Hooi, Wen Liu, Takashi Kadowaki, Masaomi Nangaku, Adeera Levin, David Z I Cherney, Aldo P Maggioni, Roberto Pontremoli, Rajat Deo, Shinya Goto, Xavier Rossello, Katherine R Tuttle, Dominik Steubl, Dan Massey, Martina Brueckmann, Martin J Landray, Colin Baigent, Richard Haynes
PMID: 39453837
Introduction
Sodium-glucose co-transporter-2 inhibitors (SGLT2i or Flozins) have been correctly called ‘the statins of the 21st century’. (Braunwald E, Eur Heart J 2022). Multiple trials have been conducted where improved outcomes with SGLT2 inhibitors have been demonstrated in patients with diabetes and high cardiovascular risk (EMPA-REG OUTCOME, CANVAS, SCORED, VERTIS, DECLARE-TIMI 58), diabetes and chronic kidney disease (CKD) (CREDENCE), proteinuric CKD with or without diabetes down to eGFR 25 ml/min (DAPA-CKD), heart failure with reduced ejection fraction (DAPA-HF, EMPEROR-Reduced), and heart failure with preserved ejection fraction (DELIVER, EMPEROR-Preserved).
SGLT2 Inhibitors and Renal Outcomes: A comparison of RCTs
Infographic by Priti Meena, from EmpaKidney summary
The EMPA-KIDNEY trial was designed to observe the effect of empagliflozin in a wide variety of patients with CKD at increased risk of progression to end stage kidney disease (ESKD), with or without diabetes and/or albuminuria. It was initiated based on the findings of the EMPA-REG OUTCOME trial (Herrington et al, CKJ 2018) where a subgroup analysis of patients with diabetes and established cardiovascular disease with eGFR 30-60 ml/min (Wanner et al, Circulation 2018) revealed that the cardioprotective effects are sustained in spite of the decreasing glucose lowering effect of the flozins. Various mechanisms have been implicated which contribute to improvement in the cardiovascular outcomes including natriuresis, diuresis, and improvement in BP (Griffin et al, Circulation 2020; NephJC summary).
EMPA-KIDNEY was a multinational, randomized, double-blind, placebo-controlled trial of empagliflozin. It was conducted in 241 centers in 8 countries in Europe, North America, and East Asia. The trial was stopped early as it met the efficacy outcomes at the interim analysis. The key results from the trial were:
There was a 28% risk reduction of the primary outcome of kidney disease progression or cardiovascular deaths in the empagliflozin group compared to the placebo group (HR = 0.72, 95% CI = 0.64 to 0.82).
There was a 14% relative risk reduction in the hospitalization due to any cause (HR = 0.86, 95% CI = 0.78 to 0.95).
A significant reduction in the rates of kidney disease progression was noted with relative risk reduction of 29%.
The effect of empagliflozin across the prespecified subgroups (DM, GFR and uACR) was mostly consistent.
The annual rate of eGFR decline was slower in the empagliflozin arm by 0.75 ml/min despite the initial decline in eGFR in the first 2 months of randomization. This was preserved in the key subgroups including those with uACR < 300 mg/g.
The drug was well tolerated with no significant increase in hyperkalemia, serious urinary tract infections or AKI.
The empagliflozin arm was associated with a mean difference in body weight of −0.9 ± 0.1 kg, systolic BP of −2.6 ± 0.3 mm Hg, and HbA1c of −0.39 mmol/mol. The geometric mean uCAR was 19% lower in the empagliflozin group (95% CI 15 to 23).
The benefits of empagliflozin were also evident in the non-diabetic populations (confirming DAPA-CKD) as well, and can be used till a GFR of 20.
EMPA-KIDNEY showed that empagliflozin significantly reduces the risk of progression of kidney disease and cardiovascular death, without a significant increase in adverse events. The medication’s beneficial effect was preserved across different eGFRs among patients with and without diabetes, and was greater amongst those with moderately increased albuminuria. This helped to establish that it may never be too late to initiate SGLT2i agents. A unique feature of EMPA-KIDNEY trial was including 3192 patients (48%) with uACR<300 mg/g (<34 mg/mmol), with 1328 being normoalbuminuric. Furthermore, empagliflozin was used till eGFR of 20 ml/min in the trial. It is a known fact that flozins have cardiovascular benefits among patients with diabetes who do not have albuminuria, based on the cardiovascular outcome trials in diabetes (EMPAREG, CANVAS, DECLARE). However, do the cardiovascular benefits persist even after the cessation of the drug?
Why was there a need in the follow up trial? It is always nice to revisit trials to see if the future is just as bright as the controlled environment for study participants.
The Post-Trial Study
Methods
This was a post trial follow-up study of EMPA-KIDNEY. EMPA-KIDNEY was a multinational, randomized, parallel-group, double-blind, placebo-controlled trial of the SGLT2 inhibitor empagliflozin (details here and in the image below).
The post trial follow up was an optional substudy which was conducted at 185 of the 241 (77%) trial centers in 7 of the original 8 countries. All surviving patients from these participating centers were eligible for post trial follow-up.
Patients: The EMPA-KIDNEY eligible patients were adults with the race-adjusted eGFR (calculated with CKD-EPI 2009 formula)
At least 20 but less than 45 ml/min/1.73 m2, regardless of the level of albuminuria, or
At least 45 but less than 90 ml/min/1.73 m2 with a uACR of at least 200 mg of albumin per gram of creatinine at the screening visit for the active trial
The patients were also required to be taking a clinically appropriate dose of a single agent renin angiotensin system inhibitor (RASi) unless contraindicated.
Inclusion and Exclusion Criteria from EMPA-KIDNEY
Procedures: All unused doses of trial empagliflozin or placebo drugs were collected from the patients at the final follow up visit of the active trial. The local practitioners were informed of the conclusions of the trial. They were at liberty to prescribe open-label SGLT2 inhibitors (empagliflozin) to the patients when indicated and conduct routine follow up of the patients as indicated. Post trial follow up was conducted to assess additional efficacy and cause specific outcome data with respect to mortality. Medical record review was the main method of follow up, which was supplemented with registry data in the United Kingdom and Malaysia. Six-monthly reviews were conducted where data on patients vital status, current kidney replacement status, recent blood creatinine or any current use of relevant medications (SGLT2i, RASi, and mineralocorticoid receptor antagonists, MRA) were collected by a custom-made information technology system.
Outcomes: The primary outcomes were composite kidney disease progression (sustained reduction in eGFR of less than 40%, development of ESKD, sustained eGFR <10 ml/min/1.73 m2, or death from kidney failure) or cardiovascular death. The secondary outcomes were kidney disease progression alone, composite of ESKD or any cause of death or ESKD alone. The tertiary outcomes were death from any cause and death from cardiovascular and non-cardiovascular causes. The subgroup analysis was done according to diabetes status, eGFR, uACR, and primary cause of kidney disease. eGFR assessment during follow up required two consecutive measurements at least 30 days apart. The analyses assessed yearly changes and varied eGFR measurement windows. The figure below demonstrates the possible scenarios for post trial results.
Statistical analysis: The authors used prespecified Cox proportional hazards regression models that included adjustment for categorized baseline variables specified in the minimization algorithm (age, sex, previously diagnosed diabetes, eGFR, urinary albumin to-creatinine ratio, and geographic region) to estimate hazard ratios and 95% confidence intervals for empagliflozin, as compared with placebo, in time-to-event analyses. The authors calculated Kaplan–Meier estimates for the time until each of the primary- and secondary outcome events. Mixed-model repeated-measures approaches were used to estimate the mean eGFR at each follow-up time point throughout the entire follow-up period on the basis of local laboratory measurements.
Funding: EMPA-KIDNEY was sponsored by Boehringer Ingelheim, the manufacturer of empagliflozin, and Eli Lilly through a grant provided to the University of Oxford.
Results
Study Population
Between May 2019 and April 2021, the EMPA-KIDNEY trial enrolled 6,609 patients with a median active trial duration of 2.0 years. Among the 6,253 patients who completed the active trial phase, 1,362 participants (21.8%) were excluded from the post-trial study due to either not providing consent or site-specific logistical limitations, which included all Japanese study locations. During the post-trial follow-up period, 4,891 patients were followed up for a median of 2 years. Of those, 86 patients (1.8%) had missing vital status information, and 7 patients (0.1%) withdrew their consent during this post-trial follow up period (PTFU).
Figure S1. Trial profile for active and post-trial follow up periods. From Herrington et al, NEJM 2024.
In the post-trial follow-up cohort, mean age was 63±14 years, with 34% woman and 56.9% non-diabetic. Patients had a mean eGFR of 37±14 ml/min/1.73 m2 and median urinary albumin-to-creatinine ratio of 317 mg/g. Most patients (71.3%) had non-diabetic kidney disease, and 49.1% had albumin-to-creatinine ratios >300 mg/g. Finally, 62.4% of patients identified as white. Compared to non-participants, follow-up participants were less likely to be Asian, slightly younger, had lower eGFR and albumin-to-creatinine ratios, but faced higher kidney failure risk, with the Tangri KFRE equation.
Table 1. Characteristics of the Post-Trial Study Patients at Randomization. From Herrington et al, NEJM 2024.
Table 1 (con’t). Characteristics of the Post-Trial Study Patients at Randomization. From Herrington et al, NEJM 2024.
Table S2. Baseline characteristics at randomization for participants in post-trial follow-up (PTFU) and those not in PTFU. From Herrington et al, NEJM 2024.
During the active-trial period, empagliflozin (or an open-label SGLT2i) was administered to 90% of patients in the empagliflozin group, while only 2% of the placebo group received an open-label SGLT2i. In the post-trial phase at 12 months, after excluding patients on kidney-replacement therapy, open-label SGLT2 inhibitors were prescribed to 43.3% (880 of 2032) of the empagliflozin group and 40.7% (797 of 1960) of the placebo group. By 24 months, these numbers increased to 50.2% in the empagliflozin group and 47.7% in the placebo group. Patients who did not start an open-label SGLT2i during the post trial period were more likely to be from Asia, were less likely to have previously diagnosed diabetes, had a lower eGFR, had a notably higher risk of kidney failure, and were less likely to be receiving RASi.
At the start of post-trial-follow up, 86% of participants were on RASi. The average use of RASi declined over time, but remained similar in both groups (approximately 68%).
Table S4. Use of RAS inhibitors and mineralocorticoid receptor antagonists over time. From Herrington et al, NEJM 2024.
Table 2. Use of SGLT2 inhibitors. From Herrington et al, NEJM 2024.
Primary and Secondary Outcomes
In the combined active- and post-trial periods, empagliflozin demonstrated superior outcomes with 26.2% (865 of 3304) of patients experiencing kidney disease progression or cardiovascular death (the primary outcome) compared to 30.3% (1001 0f 3305) in the placebo group. The strongest protective effects occurred during the active trial (HR 0.72; 95% CI, 0.64 to 0.82) and early post-trial phase, especially within the first 6 months (HR 0.60; 95% CI, 0.38 to 0.93). The post-trial benefit decreased on further follow up during the first year (HR 0.76; 95% CI, 0.6 to 0.96), and further in the second year (HR 0.90; 95% CI, 0.75 to 1.07).
Figure 1. Progression of kidney disease or death from cardiovascular causes. From Herrington et al, NEJM 2024.
Table 3. Primary, secondary and tertiary outcomes during the combined active and post-trial periods. From Herrington et al, NEJM 2024.
Empagliflozin reduced the risk of kidney disease progression by 21% compared to placebo (23.5% vs 27.1%) during the combined trial periods (HR 0.79; 95% CI, 0.72-0.87). The medication also lowered the risk of ESKD to 9% compared to 11.3% in the placebo group (HR 0.74; 95% CI, 0.64 to 0.87). The combined risk of ESKD or death from any cause was reduced to 16.9% in the empagliflozin group versus 19.6% in the placebo group (HR 0.81; 95% CI, 0.72 to 0.90). During the post-trial period, there was reduced risk for kidney disease progression (HR 0.89; 95% CI, 0.77 to 1.02) and ESKD (HR 0.80; 95% CI, 0.66 to 0.98), respectively in the empagliflozin group.
Figure S3. Effect of allocation to empagliflozin on secondary outcomes over the entirety of follow-up. From Herrington et al, NEJM 2024.
*Absolute difference in number of events per 1000 participants allocated to receive empagliflozin during the active trial period. A) KM-plots for the secondary outcomes for the entire follow-up period (active + post-trial periods combined). By contrast, B) hazard ratios in those originally allocated empagliflozin versus those originally allocated placebo separately for (a) the active trial, and (b) the post-trial period during which time no participant took study drug but some were started on non-trial SGLT2i (not necessarily empagliflozin).
Based on the Kaplan-Meier analysis, the empagliflozin group showed significant benefits compared to placebo. The absolute benefits of empagliflozin were derived from the differences between Kaplan-Meier survival curves of treatment and control groups at specific time points, expressed as events prevented per 1000 patients. At the active-trial period conclusion, the empagliflozin group had 57±14 fewer primary outcome events per 1000 patients, which remained substantial at 45±14 fewer events through the combined active and post-trial periods. Regarding ESKD, the empagliflozin group demonstrated 26±8 fewer cases per 1000 patients during the active trial, maintaining similar protection with 25±10 fewer cases through the combined period. The composite outcome of death or ESKD showed 25±11 fewer events during the active trial and 32±12 fewer events through the combined period in the empagliflozin group per 1000 patients.
Table S6. Benefits of allocation to empagliflozin on primary and secondary outcomes over the entirety of follow-up. From Herrington et al, NEJM 2024.
Plot of mean eGFR by time calculated from an MMRM model for the entirety of the trial with all analyses based on local creatinines. The vertical lines indicate the 95% confidence intervals for the estimated means.
*Use of SGLT2 inhibitor based on reporting taking at least 80% of study drug or use of open-label SGLT2 inhibitor.
Subgroup analysis
The relative effects on the primary outcome over the entirety of follow-up were consistent in subgroups split by baseline diabetes status, levels of kidney function, levels of albuminuria and primary cause of kidney disease.
Figure 2. Subgroup analysis of the primary outcome. From Herrington et al, NEJM 2024.
Tertiary analysis
During the combined trial periods, patients receiving empagliflozin showed a lower risk of death from a cardiovascular cause of 3.8% compared to 4.9% in the placebo group (HR 0.75; 95% CI: 0.59-0.95). Deaths from non-cardiovascular causes remained consistent between both groups at 5.3% (HR 0.97; 95% CI: 0.79-1.20). Overall mortality rates were 9.1% (301 deaths) in the empagliflozin group and 10.2% (336 deaths) in the placebo group (HR 0.86; 95% CI: 0.74-1.01).
The empagliflozin group had a mean final eGFR of 31.4±0.2 ml/min/1.73m2 versus 30.6±0.2 ml/min/1.73m2 in the placebo group. There was a significant difference of 0.8 ml/min/1.73m2 between the groups. The effect was consistent across all major patient subgroups analyzed.
Discussion
During the two year post-trial follow-up of EMPA KIDNEY, original treatment assignments remained blinded, both groups showed comparable but unacceptably low rates of open-label SGLT2i use. Notably, patients with CKD who were originally assigned to empagliflozin continued to show cardiorenal benefits even after stopping the study medication for the first few months to a year, and the benefit was then lost.
Key Strengths: This trial had broad eligibility criteria in which 56.9 % of the population were non diabetic and 48.9% had urine albumin creatinine ratios of 300 or less. The initial trial was limited by insufficient cardiovascular events, hampering its ability to assess secondary and tertiary cardiovascular outcomes effectively (Herrington et al, NEJM 2023). This limitation was largely overcome in the follow-up study, which allowed for detection of almost doubled the number of primary outcome events from 990 to 1,866 through extended observation periods. The study's robustness was enhanced by its large sample size, strong medication adherence, exceptional post-trial participation rates, and comprehensive follow-up data collection.
Limitations: The post-trial follow used local creatinine measurements rather than centralized testing, relied primarily on medical reviews or patient/caregiver contact for data collection (introducing potential recall bias), excluded Japanese participants (reducing Asian representation compared to the original trial), and did not evaluate adverse events. Data quality and standardization may have been compromised due to these methodological constraints.
Carry Over Effect / Legacy Effect: In this two year post-trial follow up study, carry over effects of 13% relative risk reduction for the primary outcome was less than the 28% reduction obtained while taking empagliflozin during the original 2-year active trial. Much of the post-trial benefit occurred early (in the first 6 months) and lasted only 12 months. Thus it is essential to use flozins uninterrupted, long-term for maximizing cardiorenal benefits. Flozination during the active trial period may have reduced hyperfiltration leading to preservation of kidney function and thus reduced risk of ESKD (Vallon et al, Diabetologia 2017). This may explain the carry over effects of SGLT2i in the original empagliflozin group, for the first few months after trial completion.
Long-term SGLT2i and Real World Evidence: This study reflected real-world clinical practice conditions, with local physicians managing patients' kidney function monitoring and SGLT2i prescriptions as clinically indicated. Medication usage patterns were representative of real-world practice: 43% and 40% of empagliflozin and placebo groups respectively received open-label SGLT2 inhibitors, while RAS inhibitor use was 68%, and mineralocorticoid receptor antagonists were prescribed to 6-7% of patients across both groups.
Recent real-world evidence aligns with EMPA-KIDNEY's follow-up results. A large study across twenty US healthcare systems (62,197 patients) showed empagliflozin reduced kidney outcomes by 25% and cardiovascular events by 19% compared to DPP4 inhibitors over a median 1.1 year follow up (Edmonston et al, Am J Cardiol 2024). Similarly, an Israeli database study of 63,145 person-years demonstrated that patients taking SGLT2i had 28% fewer kidney-specific outcomes and 20% reduction in kidney-or-death outcomes versus those on DPP4 inhibitors, with slower eGFR decline (0.49 ml/min/1.73 m2 annually), even in patients without pre-existing cardiovascular or kidney disease (Cohen et al, CJASN 2023). These real-world findings corroborate the sustained benefits observed in the EMPA-KIDNEY follow-up study.
Effects of De-flozination: Recent Veterans Affairs study of CKD patients (stages 3-4) found concerning discontinuation rates in CKD patients (SGLT2i discontinued, aka de-flozination, in 37% patients), with cessation linked to increased risk of death and heart failure hospitalization (Gregg et al, JASN 2024). EMPA-KIDNEY follow-up showed persistent cardiorenal benefits for at least 12 months post-discontinuation.
In contrast, the EMPEROR trials in heart failure patients, demonstrated that while empagliflozin provided sustained cardiovascular benefits during treatment (24% risk reduction), these effects quickly reversed within 30 days of discontinuation, suggesting rebound effects in sodium and water homeostasis(Packer et al, Circulation 2023).
From Evidence to Action: Bridging the Knowledge-Practice Divide
Recent studies highlight concerning trends in SGLT2i use. Despite strong clinical evidence supporting SGLT2i use in patients with CKD, regardless of diabetes status, there was an unexpectedly low adoption, with only 41.5 % of the study population receiving these drugs by their local practitioners. A 2024 Veterans Affairs study discussed earlier found that SGLT2i were stopped in 37% of patients with stage 3-4 CKD (Gregg et al, JASN 2024). The situation appears even more challenging when examining prescription patterns. Data from the Veterans Health Administration system during 2019-2020 revealed that a mere 10.7% of patients with type 2 diabetes received SGLT2i prescriptions (Lamprea-Montealegre et al, JAMA 2022). US CDC data further exposed significant demographic disparities in the prescription of newer diabetes medications, including SGLT2i, GLP-1RA, and DPP-4 inhibitors. Even among groups with relatively higher prescription rates - older adults (21.3%), women (17.8%), and non-Hispanic White patients (20.4%) - utilization remained suboptimal (CDC. CKD in US, 2021). This pattern of slow adoption echoes the historical challenges faced with implementing RASi, despite compelling clinical trial evidence supporting their use.
This implementation gap stems from multiple barriers in CKD care including clinical inertia, low disease awareness, low kidney disease education among patients and health providers, primary care limitations, polypharmacy, multiple morbidity, disjointed care delivery, healthcare iniquities and cost constraints (Nee et al, NDT 2023). Healthcare delivery is often fragmented due to providers working in isolation, creating a "medical complexity care conundrum" where multiple specialists' overlapping responsibilities lead to delayed treatments and lack of care ownership. The solution lies in breaking down these silos and establishing interdisciplinary care models where providers, particularly nephrologists, expand their roles beyond traditional boundaries to ensure coordinated, effective patient care (Ambruso et al, KI Reports 2024). The global kidney community must adopt a comprehensive approach using the Chronic Care Model framework that integrates patient education, self-management, collaborative care, technology, and policy changes to overcome barriers in cardiorenal care and improve outcomes.
Medical complexity care conundrum. From Ambruso et al, KI Reports 2024.
Conclusion
SGLT2 inhibitors a.k.a. Flozins have emerged as an effective therapeutic tool for managing chronic kidney disease, demonstrating benefits in patients with and without diabetes. The key takeaway from this study for healthcare providers is straightforward - initiate treatment early, ensure consistent therapy, and recognize that the protective effects are not sustained after discontinuation. Flozinate early and persist with flozination.
Summary Prepared by
Kajaree Giri
Nephrologist, Hyderabad, India
Shahin Mohammed
Nephrologist, Dubai, UAE
NSMC interns Class of 2024
Reviewed by Brian Rifkin, Cristina Popa, Pallavi Prasad and Jade Teakell
Header Image created by AI, based on prompts by Evan Zeitler