#NephJC Chat
Tuesday June, 11th, 2024 at 9 pm Eastern (AEST = June 12th, 11am)
Wednesday June 12th, 2024, at 9 pm Indian Standard by Time and 3:30 pm GMT (AEST = June 13th, 2am)
N Engl J Med 2024 May 24. doi: 10.1056/NEJMoa2403347. Online ahead of print.
Effects of Semaglutide on Chronic Kidney Disease in Patients with Type 2 Diabetes
Vlado Perkovic, Katherine R Tuttle, Peter Rossing, Kenneth W Mahaffey, Johannes F E Mann, George Bakris, Florian M M Baeres, Thomas Idorn, Heidrun Bosch-Traberg, Nanna Leonora Lausvig, Richard Pratley; FLOW Trial Committees and Investigators
PMID: 38785209
Introduction
Traditionally, and rightly so, the care of patients with chronic kidney disease (CKD) is centered around approaches that slow disease progression. CKD is a relentless wildfire that sets out to engulf the land beneath a patient's (and a nephrologist's) feet in totality, sooner or later. The most common cause of CKD in the world, diabetic kidney disease (DKD), is perhaps the more daunting variety since currently no therapies offer the dramatic reversals possible for other glomerular diseases where a nephrologist, through immunosuppressives and the like, may exert some influence. We have relied upon personal experiences, expert opinions, investigations, well-charted trials and a whole gamut of pharmaceutics to improve the care of CKD patients, but we have yet to minimize our patients’ risks to the point of halting DKD.
For decades, we relied on the magic of glycemic control and control of hypertension to turn the tide of worsening DKD, only to see some patients with decent HbA1Cs still suffer from a decline of kidney function. New additions (Sodium glucose transporter 2 inhibitors [SGLT2i/Flozin] and non-steroidal mineralocorticoid antagonists [nsMRAs]) to the bedrock of RAS inhibition (RASi) covered more territory and improved outcomes. While the nephroprotective effects of SGLT2i were initially derived from secondary endpoints of CV outcome trials (CVOTs), direct confirmation came from dedicated kidney focused trials: CREDENCE (Perkovic et al NEJM 2019| NephJC Summary), DAPA-CKD (Heerspink et al NEJM 2020| NephJC Summary) and finally EMPAKIDNEY (Herrington et al NEJM 2023| NephJC Summary). With every trial, we’ve felt a step closer to diverting the surging immense CV and renal risk in the DKD population. However, even with a three medication regimen (RASi, SGLT2i, and nsMRAs), there is still undeniably room for improvement, because residual risk in diabetic still looms large.
Slide from Peter Rossing’s presentation- Current treatment strategies, unmet medical need, FLOW rationale, during FLOW session, 61st European Renal Association Congress
The 4th pillar of DKD came as from a splash from the past. The potential nephroprotective effects of GLP-1RAs had been reported in a further analysis of renal outcomes in a series of CVOTs trials that showed reduction in proteinuria and hinted at slowing eGFR decline.
Figure from Rossing et al NDT 2023
This metabolic multi-tasker does it all; stimulating glucose-dependent insulin release, suppressing glucagon production, and even slowing gastric emptying to keep blood sugar levels steady. Beyond glycemia, GLP-1RAs also have beneficial effects on CV outcomes, and produce weight loss in a safe manner. In the kidney, GLP-1RAs directly induce natriuresis and diuresis by acting on the proximal convoluted tubule (PCT) through inhibition of sodium-hydrogen exchanger 3 (NHE3). You can read more about GLP-1RAs and the kidney in NephMadness 2024 (Animal House region).
But is it a whole new therapeutic target for DKD? Does it stand tall with RASi, Flozins, and nsMRAs as the fourth pillar of CKD management in patients with diabetes? The FLOW trial (Evaluate Renal Function with Semaglutide Once Weekly) is the first designed to assess the effect of a GLP-1RA on primary composite kidney outcomes among patients with CKD and type 2 diabetes (T2DM).
Figure from van der Aart-van der Beek AB, Nat Rev Nephrol, 2020
The Study
Methods
This was an international, event-driven, double-blind, parallel-group, randomized, placebo-controlled trial conducted at 387 sites in 28 countries comparing semaglutide to placebo in patients with T2DM and CKD, both being treated with baseline standard of care.
Figure from Rossing et al NDT 2023
Study Population
Inclusion criteria
Adult
T2DM (HbA1c ≤10%)
High risk of CKD defined (using Scr & CKD-EPI 2009 equation) as:
eGFR ≥50 to ≤75 ml/min/1.73m2 + UACR >300 to <5000 mg/g (~ 34 to 565 mg/mmol)OR
eGFR ≥25 to <50 ml/min/1.73m2 + UACR >100 to <5000 mg/g (~ 11 to 565 mg/mmol)
Receiving stable (4 weeks prior to enrollment) maximum tolerated dose of a RASi (unless contraindicated or not tolerated)
Exclusion Criteria
Congenital or hereditary kidney disease (including autoimmune kidney disease or polycystic disease)
Use of any GLP1-RA within 30 days prior to screening
MI, stroke, hospitalization for unstable angina or TIA within 60 days prior to screening
Current NYHA Class IV heart failure
Planned coronary, carotid or peripheral artery revascularization
Current (or within 90 days) chronic or intermittent HD or PD
Uncontrolled proliferative diabetic retinopathy or maculopathy*
Combination use of ACEi and ARB
Presence or history of malignant neoplasm within 5 years prior to the day of screening
Prior solid organ transplant or awaiting solid organ transplant
Pregnancy and/or breastfeeding
*For excluding people with proliferative retinopathy, results of an eye examination (fundus photography, or slit lamp biomicroscopy) performed by an ophthalmologist or optometrist had to be available. This exclusion was based upon findings in rodent studies. Also note that there was no obesity or BMI eligibility criterion, and though the highest HbA1c allowed was 10%, there was no lower limit of glycemia control for exclusion.
Interventions
Eligible patients were randomized (1:1) to either receive subcutaneous semaglutide (starting at 0.25 mg weekly and increased at 4 week intervals to a maintenance dose of 1 mg weekly) or placebo. SGLT2i and MRA use was permitted, and randomization was stratified by use of flozins at baseline. Participants with GFR > 60ml/min/1.73m² were capped at 20%. Glycemic control was encouraged with diet and exercise per discretion of site/physician. Injection training was repeated at week 4, week 26, and yearly.
Figure from Rossing et al NDT 2023
Primary Outcome
This was a composite of time of first occurrence of:
Onset of kidney failure (persistent eGFR <15 ml/min 1.73m² for >28 days or initiation of long-term KRT, kidney transplantation)
Persistent ≥50% reduction in eGFR from baseline
Death from kidney or CV causes
Secondary Outcomes
Confirmatory secondary outcomes were analyzed hierarchically. If the primary outcome showed superiority, secondary outcomes were tested in a pre-specified order to control the type I errors
Total eGFR slope (the annual rate of change in eGFR from randomization to the end of the trial)
Major cardiovascular events (a composite of nonfatal MI, nonfatal stroke, or death from CV causes)
Death from any cause
The other supportive secondary outcomes are shown in the figure below.
Slide from FLOW session, 61st European Renal Association Congress
Sample Size and Analytic Plan
This trial was event-driven, with a minimum of 854 primary outcome events (including 515 kidney events) which would provide 90% power to detect a 20% risk reduction. Based on a ~ 8.5% annual event rate in the placebo arm, this would require about 3508 participants to be enrolled. The trial used an intention-to-treat principle, including all participants who underwent randomization, regardless of their adherence or changes in medication. An interim analysis was planned after two thirds of the total events had accrued with clear stopping rules and an alpha spending function to account for this interim look (in fact, the interim analysis did not take place until 741 events, out of the planned 854, took place). Time from randomization to first composite kidney event, first composite major cardio-vascular event, and first all-cause death event were analyzed using Cox proportional hazards model, with flozins at baseline used for stratification. The eGFR slope was analyzed with linear mixed effects.
For the primary composite outcome and secondary hierarchical outcomes, the threshold for two-sided statistical significance was set to 0.0322 (one-sided level 0.00161) using Lan-DeMets alpha spending function, which allows to adjust significance levels as data were reviewed for the interim analysis, reducing the chance of false positive results.
Funding
Funding was provided by Novo Nordisk which manufactures semaglutide. The sponsor also managed trial operations and conducted the analyses which were independently verified with the use of the original data by Statogen Consulting. The first author wrote the first draft of the manuscript, and all the authors (four of whom are Novo Nordisk employees) contributed to subsequent revisions. Technical editorial assistance was provided by OpenHealth and funded by the sponsor.
FLOW trial management, from Rossing et al, NDT 2023
Results
Interim analysis
The trial was halted early in October 2023 after the DSMB performed an interim analysis after 570 events had accrued, and stopping rules were met. Mean follow-up duration was 3.4 years (IQR = 0-4.5 years). A total of 3472 (98.6%) patients completed follow-up or died during this period. By the time all follow up visits were completed, 741 events had occurred (close to the 854 initially planned, and thus ~ 170 more after interim analysis performed).
Study Population
A total of 5581 participants from 28 countries were screened from June 2019 through May 2021, of whom 3533 were randomized. Drug discontinuation occurred in 26% of the participants, slightly higher for semaglutide (12.6%) than placebo (11.3%) mostly driven by GI effects. Adherence to the trial regimen averaged 89% of the planned time during the study period.
Figure S1. Flow of patients through the trial Perkovic et al, NEJM 2024
The randomized participants had a mean age of 66.6 years. Men represented about two-thirds of the trial population. There were patients enrolled from Asia/Europe/North America, but very few Black participants (4.5%) overall. However, if one examined the participants only from North America, Black patients represented about 18% (160/865). The mean BMI ~ 32 kg/m² (10.8% with BMI >40 and 11.6% with BMI <25). The mean eGFR was 47 ml/min/1.73m2, and about two-thirds were CKD G3, two-thirds had macroalbuminuria and a third had microalbuminuria. The mean HbA1c was 7.8%, with about half of patients having an HbA1c ≥ 7.5%, and 61% were on insulin at baseline. In terms of CV disease, 20% had a history of chronic heart failure and 22% had previous MI or stroke. RASi use was almost universal (60% on ARB and 35% on ACEi), but only 15% were on a SGLT2i at baseline. On the other hand, statin use was appropriately high at 80%.
Table 1: Baseline characteristics. Perkovic et al, NEJM 2024
Figure from Rossing et al, NDT 2023
Distribution of study population by eGFR and albuminuria. Adapted from Supplementary Table S2. Perkovic et al, NEJM 2024
Primary Outcome
The primary outcome occurred in 331 patients (5.8 per 100 patient-years) in the semaglutide group vs 410 patients in the placebo group (7.5 per 100 patient-years) with a 24% risk reduction with semaglutide (HR 0.76, 95% CI, 0.66 to 0.88; p = 0.0003).
A composite of the kidney-specific components of the primary outcome was also lower in the semaglutide group (HR 0.79; 95% CI, 0.66 to 0.94).
Figure 1A, 1B: Primary outcome. Perkovic et al, NEJM 2024
Secondary Outcomes
eGFR Slope
The mean annual decrease in the eGFR (total slope) was significantly lower in the semaglutide group by 1.16 ml per minute per 1.73 m2 (95% CI, 0.86 to 1.47; p<0.001). Note the acute drop in GFR followed by a U-shaped rebound in the semaglutide arm. At week 12 (a.k.a. acute slope), the difference between the two groups was minimal (-0.03 ml/min/1.73m2). From week 12 to the end of the trial (a.k.a. chronic slope) there was a significant difference (0.94 ml/min/1.73m2). Also note that after 104 weeks or so, the placebo group flattens on the eGFR slope progression, possibly due to survival bias after the patients who started dialysis or had events (reached primary endpoint) dropped out of evaluation. Interestingly enough, the total slope using cystatin C was also significantly slower for semaglutide versus placebo (3.39 ml/min/1.73m2).
Figure 1D from Perkovic et al, NEJM 2024
CV and All-Cause Mortality
There was an 18% lower risk of CV events, a 29% lower risk of CV death and a 20% lower risk of death from any cause in the semaglutide group compared to placebo.
Figure 1C,1F, 1E from Perkovic et al, NEJM 2024
Kiney specific component outcomes
A deeper look at the individual kidney outcome components of the primary outcome was consistent with primary analysis though p values are not reported beyond the primary and key secondary outcomes as planned.
Table 2 Efficacy outcomes. Perkovic et al, NEJM 2024
Intermediate outcomes
The semaglutide group showed greater reduction in UACR by 38%, body weight by 4.1kg, HbA1C by 0.81%, and SBP by 2.23 mmHg while DBP was lower in the placebo group.
Other efficacy outcomes Figure S2. Perkovic et al, NEJM 2024
Subgroup Analyses
The primary outcome was consistent in all subgroups.
Table 3 Subgroup Analysis of the Primary Outcome. Perkovic et al, NEJM 2024
Safety Outcomes
The total incidence of serious adverse events was fewer in the semaglutide group than in the placebo group (49.6% vs. 53.8%). We can note fewer CV adverse events (in particular heart failure) in the semaglutide group and comparable incidence of AKI and severe hypoglycemia. Malignant tumors were 1% higher in semaglutide (given concern of increased thyroid cancers reported in the past [Wang et al, Endo Res 2022], and hence the exclusion of individuals with active or recent cancer). Permanent discontinuation of the drug due to adverse events occurred in 13% in semaglutide arm which was mainly due to gastrointestinal disorders (the full gamut from nausea, vomiting, diarrhea, constipation) and some more from decreased appetite and weight loss. Interestingly, the placebo group also reported discontinuations in 11.9% of participants (see table S5 for complete list). Approximately 20% of adverse events in both groups were attributed to COVID infections (slightly higher in the placebo group).
Table 3 Safety outcomes. Perkovic et al, NEJM 2024
Discussion
Semaglutide seems worthy of the overFLOWing hype. There was a significant benefit in terms of the primary outcome, with good safety and tolerability. FLOW is the first trial looking at the primary renoprotective effects of a GLP1- RA with benefits that extend down to a eGFR of 25 ml/min/1.73m2. A mentionable add-on compared to other landmark cardio-renal trials is that the FLOW trial included an older, higher-risk population: 93% with high and very high risk, 80% on lipid-lowering drugs. Being conducted in a diverse population from wider geographical areas, generalizability of results is expected to be better overall.
Slide from Cristoph Wanner’s presentation- FLOW study. Independent commentary, during FLOW session, 61st European Renal Association Congress
Like all studies FLOW does have some limitations. Since type 1 diabetics were excluded, questions still remain if this can be applied to a majority of patients <18 years. Also, a low proportion of Black patients (<5%) may limit generalizability in such a population. Lastly, SGLT2i therapy was not a requirement, as they were not the standard during the time of this trial’s enrollment/intervention. However, there were a larger number of “drop in” (initiation of SGLT2i during the study) patients in the placebo group, presumably due to elevated albuminuria and hyperglycemia, making the benefits of semaglutide even more impressive as the comparator placebo group was increasing on RASi/SGLT2i combination therapy.
Many areas of study and questions still exist. It is still a mystery how GLP1-RA mechanistically protects kidney function. Does the nephroprotective effect mostly stem from modifying the risk factors and metabolic profile related to DKD? Or could it benefit non-diabetic CKD with obesity and/or albuminuria also? Though there was no BMI cutoff in FLOW, the mean BMI was 31 - would semaglutide have a benefit in a purely non-obese population, say below BMI 25 (in the subgroup analysis the benefit seems consistent above/below BMI 30)? Can a higher dose of semaglutide, as used in CV outcome trials (e.g. SELECT, Lincoff et al NEJM 2023), strengthen this fourth pillar of treatment of DKD? Can it be used safely and effectively throughout the lifecycle of DKD, for instance in patients who commence RRT or in kidney transplant recipients, like its counterpart- the flozins? (NephJC summary)
Is this a class effect?
Different members of the GLP-1RA family have shown proven benefit for kidney outcomes in CVOTs. However, most of these trials included early stage CKD patients. We look forward to more trials that can further expand the evidence about this class of drugs. The potency for weight loss does vary by agents, so it is quite possible that (unlike flozins) the GLP1RA renal benefits may not be a simple class effect.
When and how should GLP1-RAs be added to existing therapies for DKD?
The progression and titration of the four pillars of DKD treatment (RASi, nsMRA, SGLT2i/flozins, and GLP1-RA) will require frequent monitoring. The best time and order to initiate these medications will likely need to be individualized, but further guidance and structure would help nephrologists not suffer from “therapeutic inertia”. The in press SMART-C meta-analysis presented during ERA 61th Congress, including over 3000 patients with both SGLT2i and GLP1-RA, showed no supplementary safety concerns. However, this meta-analysis included major RCTs of SGLT2is, but did not address the question of combined initiation of SGLT2i and GLP1-RA (important limitations: small proportion with GLP1-RA, patients on double therapy had more CV comorbidities, and it did not account for different GLP1-RA classes). In terms of palpable benefits, according to the analysis including flozin and nsMRA RCTs, event-free survival with SGLT2i + nsMRA + GLP1-RA was 3.2 years (CI 2.1-4.3), in favor of combined therapy versus standard of care (21.1 years vs 17.9 years). (Neuen et al, Circulation 2024)
Figure 2 from Neuen et al, Circulation 2024
Do GLP1-RAs work in non-diabetic CKD?
The SELECT trial (Lincoff et al, NEJM 2023) was a randomized controlled, double-blind, event-driven, superiority trial that included patients with a medium BMI of 27 kg/m2, but without diabetes. A pre-specified analysis showed that composite kidney outcome was lower with semaglutide 2.4 mg (1.8% vs 2.2%), HR= 0.78; 95% CI, 0.63-0.96. The treatment benefit at 104 weeks for eGFR was 0.75 ml/min/1.73m2 (95% CI 0.43-1.06; P < 0.001) in favor of semaglutide (Colhoun et al, Nat Med 2024). However, only approx. 11% of the total trial population had an eGFR <60, and >85% had UACR <30 mg/g. A mentionable possible confounder is weight loss in semaglutide arm was -9.39% (vs placebo -0.88%). The SELECT trial didn’t measure cystatin C-based eGFR, to distinguish eGFR changes due to muscle mass.
Figure 1 from Colhoun, et al, Nat Med 2024
Even if semaglutide saves lives, hearts, and kidneys, and FLOW is a global trial, implementation of the so-called fourth pillar remains a local challenge. A registry study in 2020 in the US showed only 40% of the patients with CKD were on ACEi/ ARBs, and smaller similar percentages (around 6%) for SGLT2i and GLP1-RA use (Nicholas, et al, Diabetes Obes Metab, 2023). Much has changed since 2020, but much has not. In the end, the drugs can only benefit the patients who are prescribed them and can take them.
Slide from Katherine Tuttle’s presentation, during FLOW session, 61st European Renal Association Congress
Conclusion
The FLOW trial showed that semaglutide significantly decreased the risk of progression of kidney disease in T2DM, without significant increases in serious adverse events. The beneficial effects were consistent across different eGFRs and degrees of albuminuria. Have GLP1RAs managed to turn the clock back on the natural history of diabetes and its complications? GLP-1RAs may be the game changers for CKD in the next decades. Let’s glipinate!
