ORIGINAL RESEARCH ARTICLE
`Effect of the Glucagon-Like Peptide-1 Receptor
`Agonists Semaglutide and Liraglutide on Kidney
`Outcomes in Patients With Type 2 Diabetes:
`Pooled Analysis of SUSTAIN 6 and LEADER
`
`, MD; John B. Buse , MD; Thomas Idorn, MD;
`Ahmed M. Shaman, PhD; Stephen C. Bain, MD; George L. Bakris
`Kenneth W. Mahaffey, MD; Johannes F.E. Mann, MD; Michael A. Nauck, MD; Søren Rasmussen, PhD;
`Peter Rossing, MD; Benjamin Wolthers, MD; Bernard Zinman, MD; Vlado Perkovic , PhD
`
`BACKGROUND: We assessed the effect of once-weekly semaglutide and once-daily liraglutide on kidney outcomes in type 2
`diabetes.
`
`METHODS: Pooled (n=12 637) and by-trial data from SUSTAIN 6 (Trial to Evaluate Cardiovascular and Other Long-Term Outcomes
`With Semaglutide in Subjects With Type 2 Diabetes; n=3297) and LEADER (Liraglutide Effect and Action in Diabetes: Evaluation
`of Cardiovascular Outcome Results; n=9340) were assessed for albuminuria change, annual slope of estimated glomerular
`filtration rate (eGFR) change, and time to persistent eGFR reduction (30%, 40%, 50%, and 57%) from baseline.
`
`RESULTS: The median follow-up durations were 2.1 years for SUSTAIN 6 and 3.8 years for LEADER. In the pooled analysis,
`semaglutide/liraglutide lowered albuminuria from baseline to 2 years after randomization by 24% versus placebo (95% CI,
`20%–27%; P<0.001). Significant reductions were also observed in by-trial data analyses (P<0.001 for all), the largest being
`with semaglutide 1.0 mg (33% [95% CI, 24%–40%]; P<0.001) at 2 years. With semaglutide 1.0 mg and liraglutide, eGFR
`slope decline was significantly slowed by 0.87 and 0.26 mL/min/1.73 m2/y (P<0.0001 and P<0.001), respectively, versus
`placebo. Effects appeared larger in patients with baseline eGFR <60 versus ≥60 mL/min/1.73 m2 (Pinteraction=0.06 and 0.008 for
`semaglutide 1.0 mg and liraglutide, respectively). Semaglutide/liraglutide significantly lowered risk of persistent 40% and 50%
`eGFR reductions versus placebo (hazard ratio [HR], 0.86 [95% CI, 0.75–0.99]; P=0.039 and HR, 0.80 [95% CI, 0.66–0.97];
`P=0.023, respectively). Similar, nonsignificant, directional results were observed for 30% and 57% eGFR reductions (HR,
`0.92 [95% CI, 0.84–1.02]; P=0.10 and HR, 0.89 [95% CI, 0.69–1.13]; P=0.34). In patients with baseline eGFR 30 to
`<60 mL/min/1.73 m2, the likelihood of persistent reduction for all thresholds was increased, ranging from HR 0.71 for 30% reduction
`(95% CI, 0.59–0.85; P=0.0003, Pinteraction=0.017) to 0.54 for 57% reduction (95% CI, 0.36–0.81; P=0.003, Pinteraction=0.035).
`CONCLUSIONS: In patients with type 2 diabetes, semaglutide/liraglutide offered kidney-protective effects, which appeared
`more pronounced in patients with preexisting chronic kidney disease.
`
`Key Words: albuminuria ◼ chronic kidney disease ◼ eGFR ◼ glucagon-like peptide-1 receptor agonists ◼ liraglutide ◼ semaglutide ◼ type 2 diabetes
`
`Type 2 diabetes (T2D) is a major risk factor for the
`
`development and progression of chronic kidney
`disease, commonly referred to as diabetic kidney
`
`disease (DKD).1 Approximately 50% of people with T2D
`will develop DKD in their lifetime, and approximately half
`of kidney failure cases are ascribed to diabetes.1,2 As the
`
`
`Correspondence to: Vlado Perkovic, PhD, The George Institute for Global Health, The University of New South Wales, Level 5, 1 King Street, Newtown NSW 2042,
`Australia. Email vperkovic@georgeinstitute.org.au
`Supplemental Material, the podcast, and transcript are available with this article at https://www.ahajournals.org/doi/suppl/10.1161/CIRCULATIONAHA.121.055459.
`For Sources of Funding and Disclosures, see page 584.
`© 2021 The Authors. Circulation is published on behalf of the American Heart Association, Inc., by Wolters Kluwer Health, Inc. This is an open access article under the
`terms of the Creative Commons Attribution License, which permits use, distribution, and reproduction in any medium, provided that the original work is properly cited.
`
`Circulation. 2022;145:575–585. DOI: 10.1161/CIRCULATIONAHA.121.055459
`
`February 22, 2022
`
`575
`
`Circulation
`
`Circulation is available at www.ahajournals.org/journal/circ
`
`Novo Nordisk Exhibit 2015
`Mylan Pharms. Inc. v. Novo Nordisk A/S
`IPR2023-00724
`Page 00001
`
`

`

`DKD
`eGFR
`GLP-1 RA
`
`HbA1c
`HR
`LEADER
`
`Nonstandard Abbreviations and Acronyms
`CKD-EPI
` Chronic Kidney Disease Epidemiol-
`ogy Collaboration
`diabetic kidney disease
`estimated glomerular filtration rate
` glucagon-like peptide-1 receptor
`agonist
`glycohemoglobin
`hazard ratio
` Liraglutide Effect and Action in Dia-
`betes: Evaluation of Cardiovascular
`Outcome Results
`modification of diet in renal disease
` Trial to Evaluate Cardiovascular and
`Other Long-Term Outcomes With
`Semaglutide in Subjects With Type 2
`Diabetes
`type 2 diabetes
`urinary albumin-to-creatinine ratio
`
`MDRD
`SUSTAIN 6
`
`T2D
`UACR
`
`Clinical Perspective
`What Is New?
`• Data suggest a kidney-protective effect with gluca-
`gon-like peptide-1 receptor agonists driven primar-
`ily by beneficial albuminuria outcomes, but definitive
`data on more severe kidney outcomes are lacking.
`• In this analysis of SUSTAIN 6 (Trial to Evaluate Car-
`diovascular and Other Long-Term Outcomes With
`Semaglutide in Subjects With Type 2 Diabetes) and
`LEADER (Liraglutide Effect and Action in Diabetes:
`Evaluation of Cardiovascular Outcome Results),
`we demonstrate the benefit of using once-weekly
`semaglutide and once-daily liraglutide on a number
`of clinically important kidney outcomes: changes in
`albuminuria, annual slope of estimated glomerular
`filtration rate change, time to persistent proportional
`estimated glomerular filtration rate reductions of
`40% and 50% from baseline, and a composite end
`point (time from randomization to first occurrence of
`kidney failure/death or proportional estimated glo-
`merular filtration rate decline).
`
`What Are the Clinical Implications?
`• The results of this study suggest that the glucagon-
`like peptide-1 receptor agonists semaglutide and
`liraglutide may add to the kidney-protective treat-
`ment options available to people with type 2 diabe-
`tes and diabetic kidney disease.
`• Further studies are required to investigate the full
`effect of glucagon-like peptide-1 receptor agonists
`on primary kidney end points in dedicated trials in
`diabetic kidney disease. SUSTAIN 6 and LEADER
`examined kidney end points as secondary outcomes
`and enrolled a majority of patients without diabetic
`kidney disease but with cardiovascular disease.
`
`global burden of T2D is increasing,3 so is the prevalence
`of kidney failure. The prevalence of kidney failure is pre-
`dicted to exceed 5 million people by 2030.4
`People with DKD have a high risk of cardiovascular
`morbidity and mortality and generate increased costs
`associated with their treatment.5 A limited number of
`treatments have been shown to lower this risk. There-
`fore, it is important to identify treatments that reduce the
`risk of cardiovascular complications and of progression
`to kidney failure in people with DKD.
`Recent developments in the management of T2D
`have identified treatments that reduce the risk of DKD
`progression. In cardiovascular outcome trials, sodium-
`glucose cotransporter-2 inhibitors and glucagon-like
`peptide-1 receptor agonists (GLP-1 RAs) have lowered
`the risk of major adverse cardiovascular events.6–12 Sec-
`ondary kidney outcomes in these cardiovascular outcome
`trials have suggested kidney benefits.6–16 In the first des-
`ignated T2D kidney outcome trials, the sodium-glucose
`cotransporter-2 inhibitors canagliflozin and dapagliflozin
`
`were shown to substantially lower the risk of hard kid-
`ney outcomes compared with placebo (patients with and
`without T2D were included in the latter trial).17,18
`This post hoc analysis aimed to evaluate the effect
`of once-weekly semaglutide and once-daily liraglutide
`therapy on a broad range of clinically important kidney
`outcomes, including changes in albuminuria, estimated
`glomerular filtration rate (eGFR) slope, and persistent
`reductions in eGFR.
`
`METHODS
`Data Sharing Statement
`Deidentified individual participant data, the study protocol,
`and a redacted clinical study report will be available accord-
`ing to Novo Nordisk data sharing commitments. The data will
`be made available permanently after research completion and
`approval of product and product use in both the European
`Union and United States. Data will be shared with bona fide
`researchers submitting a research proposal requesting access
`to data and for use as approved by the independent review
`board according to the independent review board charter. This
`and the access request proposal form and access criteria can
`be found at novonordisk-trials.com. The data will be made avail-
`able on a specialized SAS data platform.
`
`Trial Designs and Patients
`SUSTAIN 6 (Trial to Evaluate Cardiovascular and Other Long-
`Term Outcomes With Semaglutide in Subjects With Type 2
`Diabetes; URL: https://www.clinicaltrials.gov; Unique identi-
`fier: NCT01720446) and LEADER (Liraglutide Effect and
`Action in Diabetes: Evaluation of Cardiovascular Outcome
`Results; URL: https://www.clinicaltrials.gov; Unique iden-
`tifier: NCT01179048) were randomized, double-blinded,
`
`576
`
`February 22, 2022
`
`Circulation. 2022;145:575–585. DOI: 10.1161/CIRCULATIONAHA.121.055459
`
`ORIGINAL RESEARCH
`
`ARTICLE
`
`Shaman et al
`
`Semaglutide and Liraglutide Outcomes
`
`Novo Nordisk Exhibit 2015
`Mylan Pharms. Inc. v. Novo Nordisk A/S
`IPR2023-00724
`Page 00002
`
`

`

`placebo-controlled trials. Detailed methods of both trials have
`been published previously.6,7 The 2 trials included patients with
`T2D at high risk of cardiovascular events and with a glycohe-
`moglobin (HbA1c) level ≥7%. In SUSTAIN 6, patients were
`randomized to once-weekly semaglutide 0.5 mg subcutane-
`ously, once-weekly semaglutide 1.0 mg subcutaneously, or
`matching placebo for 2 years. In LEADER, patients were ran-
`domized to once-daily liraglutide up to 1.8 mg subcutaneously
`or matching placebo for 3.5 to 5 years (median 3.8 years).
`The primary outcome in both trials was major adverse cardio-
`vascular events, consisting of nonfatal myocardial infarction,
`nonfatal stroke, or cardiovascular death. Kidney end points
`were collected as secondary outcomes. Both cardiovascular
`and kidney outcomes were adjudicated by an external blinded
`event adjudication committee.6,7
`In SUSTAIN 6, serum creatinine was collected at screen-
`ing, randomization (baseline), and after baseline at weeks 2,
`4, 8, 16, 30, 44, 56, 68, 80, 92, and 104. Calculation of eGFR
`was performed using the MDRD equation (modification of diet
`in renal disease).19 Urinary albumin-to-creatinine ratio (UACR)
`was collected at baseline and after baseline at weeks 16, 30,
`44, 56, 80, and 104.
`In LEADER, serum creatinine was collected at screening,
`randomization (baseline), and after baseline at the 6-month
`visit, and then annually until final visit. Calculation of eGFR was
`performed using the MDRD equation.19 UACR was collected at
`baseline and after baseline annually until final visit.
`Urinary albumin values below the lower limit of quanti-
`fication of 3 mg/g were imputed as 1.5 mg/g (lower limit
`of quantification/2) in the post hoc calculation of UACR.
`Approximately 17% of randomized patients had UACR values
`below the lower limit of quantification in both SUSTAIN 6 and
`LEADER at baseline.
`
`Subgroups
`We evaluated outcomes for the overall pooled population and
`according to preexisting kidney disease, defined by baseline
`eGFR (using the MDRD formula) and albuminuria. Patients
`were stratified by baseline eGFR (≥90 mL/min/1.73 m2,
`60–<90 mL/min/1.73 m2, 30–<60 mL/min/1.73 m2, or <30
`mL/min/1.73 m2) and albuminuria stage (normoalbuminuria
`[UACR <30 mg/g], microalbuminuria [UACR 30−300 mg/g],
`and macroalbuminuria [UACR >300 mg/g]). Effects of sema-
`glutide and liraglutide versus placebo on albuminuria over time,
`average annual eGFR decline (slope), and eGFR at the 2-year
`visit were also evaluated in the populations treated with sema-
`glutide 0.5 and 1.0 mg and liraglutide 1.8 mg, respectively.
`
`Outcomes
`We explored the effects of semaglutide in SUSTAIN 6 and
`liraglutide in LEADER, compared with placebo, on clinically
`important kidney outcomes,20 including change in albuminuria
`from baseline to the 2-year visit (defined as UACR estimated
`ratio to baseline and presented as estimated geometric mean
`ratio between treatment and placebo), annual change in eGFR
`from baseline (total slope), and time to persistent 30%, 40%,
`50%, and 57% eGFR (57% being equivalent to doubling of
`creatinine) reductions from baseline, defined by time to first
`occurrence from randomization of the relevant reduction con-
`firmed by a subsequent measurement. If no subsequent visit
`
`was performed, the confirmation was omitted. The CKD-EPI
`equation (Chronic Kidney Disease Epidemiology Collaboration)
`was used to calculate the eGFR used in the total slope analysis.
`Other outcomes included a composite of time from ran-
`domization to first occurrence of kidney death, need for main-
`tenance kidney replacement therapy, or first occurrence of a
`reduction in eGFR of 40%, 50%, or 57% (each percentage
`analyzed separately with the first 2 components, to give 3 com-
`posite end points). We estimated total loss of eGFR over 2
`years from randomization for both trials according to treatment
`groups. We assessed these outcomes in the overall pooled
`population and in subgroups defined by UACR and eGFR at
`baseline, as defined previously. Kidney failure and death, as
`components of the original secondary end points of both trials,
`were confirmed by the event adjudication committee.6,7
`
`Statistical Analysis
`Data from all randomized patients in LEADER and SUSTAIN
`6 were included from date of randomization to the end of
`follow-up visit. In the pooled analyses, both semaglutide treat-
`ment arms and the liraglutide treatment arm were pooled and
`compared with placebo. For the pooled analyses, the trial was
`used as a fixed effect.
`Effects of semaglutide and liraglutide by trial on albumin-
`uria over time as compared with placebo were assessed using
`a mixed model for repeated measures with UACR as depen-
`dent variable (which was log-transformed owing to the nonnor-
`mal distribution), treatment and visits as fixed factors, baseline
`UACR (log-transformed) as a covariate, and the interactions
`between visits and treatment/baseline UACR, respectively.
`An unstructured covariance matrix for repeated measures
`was used. Least square means, differences, and 95% CIs
`between treatments were then back-transformed to the origi-
`nal scale. Because of the different trial durations of LEADER
`and SUSTAIN 6, treatment ratios were evaluated at 2 years
`from randomization to align comparisons of both trials. Pooled
`analyses were done using the same model with trial included
`as an additional factor.
`Slope analyses of eGFR were performed using a random-
`slope model by trial, with change from baseline as dependent
`variable, baseline value and time (in years) as covariate, treat-
`ment as a fixed factor, and the interaction between treatment
`and time. The intent-to-treat populations were used. Changes
`from baseline were assessed between each visit and used
`as the repeated measure, with the time from randomization
`as the underlying continuous time scale for the slope analy-
`ses. Patient-specific intercepts and time as random effects
`assuming a bivariate normal distribution for these effects were
`included in the model. Analyses by subgroups were performed
`by including the respective subgroups as a fixed factor and the
`interaction with treatment.
`Time-to-first-event analyses were performed using Cox
`proportional hazard models, with pooled treatment as a fixed
`factor and stratified by trial. Patients without respective events
`were censored at death or end of follow-up, whichever came
`first. Time to persistent reduction of eGFR from baseline
`(30%, 40%, 50%, and 57%) was analyzed independently from
`each other. Subgroup analyses were performed by including
`subgroup as a fixed factor and the interaction between sub-
`group and treatment. A quadratic spline Cox regression model
`
`Circulation. 2022;145:575–585. DOI: 10.1161/CIRCULATIONAHA.121.055459
`
`February 22, 2022
`
`577
`
`ORIGINAL RESEARCH
`
`ARTICLE
`
`Shaman et al
`
`Semaglutide and Liraglutide Outcomes
`
`Novo Nordisk Exhibit 2015
`Mylan Pharms. Inc. v. Novo Nordisk A/S
`IPR2023-00724
`Page 00003
`
`

`

`was also used to analyze the time to persistent reduction end
`points (30%, 40%, 50%, 57%), showing the hazard ratios
`(HRs) between the treatments (pooled GLP-1 RAs vs placebo)
`according to continuous measures.
`An I2 test was used to measure the heterogeneity between
`the 2 trials when assessing the effect of semaglutide and lira-
`glutide versus placebo on the annual eGFR slope at 2 years
`visit in the overall population and subgroups with preexisting
`DKD defined by the level of albuminuria and eGFR at baseline.
`No adjustment for multiplicity or missing values was per-
`formed. A significance level of 5% was used overall. Statistical
`analysis was performed with SAS version 9.4 (SAS Institute).
`
`Ethics
`Both studies received full approval by ethics committees or
`institutional review boards at each participating site (names
`and locations of sites are available in the supplementary infor-
`mation of the primary publications of both trials6,7) and were
`conducted according to the principles of the Declaration of
`Helsinki.21 All participants provided written informed consent
`before participation in trial-related activities.6,7
`
`Role of the Funding Source
`The sponsor participated in the study design and management,
`analysis, and interpretation of data. Three of the authors of
`this article are employees of the sponsor and, as such, were
`involved in the preparation, review, and approval of the article.
`All authors had full access to all the data in the study and had
`final responsibility for the decision to submit for publication.
`
`RESULTS
`Patients
`A total of 12 637 patients were included in the pooled
`analysis (3297 from SUSTAIN 6 and 9340 from LEAD-
`ER), with 6316 patients in the combined semaglutide/
`liraglutide group (826, 822, and 4668 patients in the
`semaglutide 0.5 mg, semaglutide 1.0 mg, and liraglutide
`1.8 mg group, respectively) and 6321 patients in the com-
`bined placebo group. The median durations of follow-up
`(after randomization) were 2.1 and 3.8 years for SUSTAIN
`6 and LEADER, respectively.6,7 Patient characteristics ac-
`cording to albuminuria and eGFR and treatment allocation
`are presented in the Table. At baseline, a total of 3063
`(24.2%) patients had an eGFR <60 mL/min/1.73 m2 and
`4726 (38.2%) patients had elevated albuminuria, either
`microalbuminuria (27.0%) or macroalbuminuria (11.2%).
`
`Effects on Albuminuria
`In the pooled analysis, semaglutide/liraglutide treatment
`lowered albuminuria from baseline to 2 years after ran-
`domization by 24% (placebo-corrected geometric mean
`ratio of relative change from baseline) compared with
`placebo (95% CI, 20%–27%; P<0.001; Figure 1). This
`was driven by data from liraglutide, given the larger num-
`ber of patients from LEADER than from SUSTAIN 6. The
`magnitude of reduction, however, was modified by base-
`
`Table. Baseline Albuminuria and eGFR Characteristics, by Treatment Group
`
`Subgroups at baseline
`
`Total population
`
`People with UACR values*
`
`Normoalbuminuria
`
`Microalbuminuria
`
`Macroalbuminuria
`
`Micro- or macroalbuminuria
`
`eGFR, mL/min/1.73 m2
`
` <60
`
` ≥60
`
` ≥90
`
` 60–<90
`
` 30–<60
`
` <30
`
`All/subtotal
`
`12 637 (100)
`
`12 375 (100)
`
`7649 (61.8)
`
`3340 (27.0)
`
`1386 (11.2)
`
`4726 (38.2)
`
`3063 (24.2)
`
`9574 (75.8)
`
`4268 (33.8)
`
`5306 (42.0)
`
`2733 (21.6)
`
`330 (2.6)
`
`Semaglutide/lira-
`glutide
`
`6316 (100)
`
`6194 (100)
`
`3842 (62.0)
`
`1695 (27.4)
`
`657 (10.6)
`
`2352 (38.0)
`
`1571 (24.9)
`
`4745 (75.1)
`
`2112 (33.4)
`
`2633 (41.7)
`
`1400 (22.2)
`
`171 (2.7)
`
` ≥60, normoalbuminuria, or both
`
`10 719 (86.6)
`
`5339 (86.2)
`
` 30–<60 and micro- or macroalbuminuria
`
`1402 (11.3)
`
` <30 and micro- or macroalbuminuria
`
` <60 and micro- or macroalbuminuria
`
`254 (2.1)
`
`1656 (13.4)
`
`723 (11.7)
`
`132 (2.1)
`
`855 (13.8)
`
`Placebo
`
`6321 (100)
`
`6181 (100)
`
`3807 (61.6)
`
`1645 (26.6)
`
`729 (11.8)
`
`2374 (38.4)
`
`1492 (23.6)
`
`4829 (76.4)
`
`2156 (34.1)
`
`2673 (42.3)
`
`1333 (21.1)
`
`159 (2.5)
`
`5380 (87.0)
`
`679 (11.0)
`
`122 (2.0)
`
`801 (13.0)
`
`The percentage of patients with urinary albumin-to-creatinine ratio (UACR) values out of the total number of patients
`was as follows: all, 97.9%; semaglutide/liraglutide, 98.1%; and placebo, 97.8%. Data are n (%). eGFR indicates estimated
`glomerular filtration rate.
`*Number of people with UACR values that were collected and analyzed at baseline was used as the denominator in calcu-
`lating the percentages in all rows, apart from the first row.
`
`578
`
`February 22, 2022
`
`Circulation. 2022;145:575–585. DOI: 10.1161/CIRCULATIONAHA.121.055459
`
`ORIGINAL RESEARCH
`
`ARTICLE
`
`Shaman et al
`
`Semaglutide and Liraglutide Outcomes
`
`Novo Nordisk Exhibit 2015
`Mylan Pharms. Inc. v. Novo Nordisk A/S
`IPR2023-00724
`Page 00004
`
`

`

`Figure 1. Effects of once-weekly semaglutide and once-daily liraglutide versus placebo on albuminuria over time.
`*Estimated geometric mean ratio calculated for each active treatment group versus the respective placebo group. Geometric mean values of
`albuminuria over time with semaglutide and liraglutide by trial as compared with placebo were estimated using a mixed model for repeated
`measures with an unstructured covariance matrix for repeated measures. Urinary albumin-to-creatinine ratio (UACR) was included as a
`dependent variable (which was log-transformed) with treatment and visits as fixed factors and baseline UACR as a covariate (log-transformed).
`Pooled analyses were done using the same model with trial also included as a fixed factor. LEADER indicates Liraglutide Effect and Action in
`Diabetes: Evaluation of Cardiovascular Outcome Results; and SUSTAIN 6, Trial to Evaluate Cardiovascular and Other Long-Term Outcomes With
`Semaglutide in Subjects With Type 2 Diabetes.
`
`line level of albuminuria: normoalbuminuria (20% [95%
`CI, 15%–25%]), microalbuminuria (31% [95% CI, 25%–
`37%]), and macroalbuminuria (19% [95% CI, 7%–30%];
`Pinteraction=0.021; Figure S1).
`Both semaglutide and liraglutide lowered albumin-
`uria compared with placebo (Figure 1). At 2 years after
`randomization, on the basis of the placebo-corrected
`geometric mean ratios of relative change from baseline,
`semaglutide 0.5 mg lowered albuminuria by 20% com-
`pared with placebo (95% CI, 10%–28%; P<0.001) and
`the 1.0 mg dose lowered albuminuria by 33% compared
`with placebo (95% CI, 24%–40%; P<0.001). At 2 years
`after randomization, albuminuria was 23% lower in lira-
`glutide-treated patients compared with placebo (95% CI,
`18%–27%; P<0.001). The effect of the semaglutide 1.0
`mg dose was statistically greater than that of liraglutide
`(P=0.024) at 2 years after randomization.
`
`Effect on eGFR Slope
`The average slope in eGFR change from baseline was
`comparable in the placebo arms of SUSTAIN 6 (average,
`−1.92 [95% CI, −2.18 to −1.67 mL/min/1.73 m2/y])
`and LEADER (average, −1.98 [95% CI, −2.10 to −1.87
`
`mL/min/1.73 m2/y]) in the overall population, and this
`was broadly consistent at different levels of baseline kid-
`ney function (Figure 2).
`In the overall population, patients randomized to sema-
`glutide 0.5 mg had a nonsignificant reduction in eGFR
`slope versus placebo (difference 0.33 mL/min/1.73
`m2/y; P=0.14), whereas randomization to semaglutide
`1.0 mg slowed kidney function loss by 0.87 mL/min/1.73
`m2/y versus placebo (P<0.0001; Figure 2). Kidney func-
`tion in patients randomized to liraglutide declined an
`average of 0.26 mL/min/1.73 m2/y slower compared
`with placebo (P<0.001).
`Baseline eGFR was found to impact the effects of
`semaglutide and liraglutide on eGFR slope, with the
`largest effect observed in patients receiving sema-
`glutide 1.0 mg with an eGFR <60 mL/min/1.73 m2/y
`at baseline. Patients in this subgroup lost 1.62 mL/
`min/1.73 m2/y less kidney function versus patients
`receiving placebo, whereas patients with baseline eGFR
`≥60 mL/min/1.73 m2/y lost 0.63 mL/min/1.73 m2/y
`less versus placebo (P<0.001; Pinteraction=0.06). The
`difference was 0.67 versus 0.21 mL/min/1.73 m2/y
`(Pinteraction=0.37) for patients randomized to semaglutide 0.5
`mg with baseline eGFR <60 or >60 mL/min/1.73 m2/y,
`
`Circulation. 2022;145:575–585. DOI: 10.1161/CIRCULATIONAHA.121.055459
`
`February 22, 2022
`
`579
`
`ORIGINAL RESEARCH
`
`ARTICLE
`
`Shaman et al
`
`Semaglutide and Liraglutide Outcomes
`
`Novo Nordisk Exhibit 2015
`Mylan Pharms. Inc. v. Novo Nordisk A/S
`IPR2023-00724
`Page 00005
`
`

`

`Figure 2. Effects of once-weekly semaglutide and once-daily liraglutide versus placebo on average annual eGFR decline.
`Effects of once-weekly semaglutide and once-daily liraglutide versus placebo on average annual estimated glomerular filtration rate (eGFR)
`decline (slope) in all patients and according to baseline eGFR. N is the number of patients whose samples/measures were available at the
`point of analysis. Slope analyses were performed on the intent-to-treat population. Slope analyses of eGFR were performed using a random-
`slope model by trial with change from baseline as dependent variable and baseline value and time (in years) as covariate and treatment as
`a fixed factor and the interaction between treatment and time. Patient-specific intercepts and time as random effects assuming a bivariate
`normal distribution for these effects were included in the model. Analyses by subgroups were performed by including the respective subgroups
`as a fixed factor and the interaction with treatment. Data shown were averaged over 2 years. LEADER indicates Liraglutide Effect and Action
`in Diabetes: Evaluation of Cardiovascular Outcome Results; and SUSTAIN 6, Trial to Evaluate Cardiovascular and Other Long-Term Outcomes
`With Semaglutide in Subjects With Type 2 Diabetes.
`
`respectively. Patients receiving liraglutide with baseline
`eGFR <60 mL/min/1.73 m2 lost 0.67 mL/min/1.73
`m2/y less kidney function than patients receiving placebo,
`whereas patients with baseline eGFR ≥60 mL/min/1.73
`lost 0.15 mL/min/1.73 m2/y (Pinteraction=0.008;
`m2
`Figure 2).
`The results were similar when the total loss of kidney
`function over 2 years after randomization of all treat-
`ments was considered (Figure S2).
`The effect of semaglutide on eGFR slope compared
`with placebo was not clearly modified by level of albu-
`minuria at baseline. Semaglutide 0.5 mg effect on eGFR
`slope compared with placebo ranged from 0.59 mL/
`min/1.73 m2 (95% CI, −0.76 to 1.94 mL/min/1.73 m2)
`in the subgroup with normoalbuminuria to 0.29 mL/
`min/1.73 m2 (95% CI, −2.16 to 2.74 mL/min/1.73 m2)
`in the subgroup with macroalbuminuria (Pinteraction=0.98)
`at 2 years after randomization. Semaglutide 1.0 mg
`effect on eGFR slope compared with placebo ranged
`
`from 1.48 mL/min/1.73 m2 (95% CI, 0.15–2.80 mL/
`min/1.73 m2) in the subgroup with normoalbuminuria
`to 2.33 mL/min/1.73 m2 (95% CI, −0.19 to 4.85 mL/
`min/1.73 m2) in the subgroup with macroalbuminuria
`(Pinteraction=0.84). The effect of liraglutide on eGFR slope
`compared with placebo appeared to be modified by the
`degree of albuminuria at baseline, with eGFR slope dif-
`ferences ranging from 0.20 mL/min/1.73 m2 (95% CI,
`−0.17 to 0.58 mL/min/1.73 m2) with normoalbumin-
`uria to 1.64 mL/min/1.73 m2 (95% CI, 0.68–2.60 mL/
`min/1.73 m2) with macroalbuminuria at 2 years after
`randomization (Pinteraction=0.023; Figure S2).
`
`Effect on Persistent Reduction in eGFR
`The overall effect of semaglutide/liraglutide (pooled)
`versus placebo on the risk of persistent reductions in
`eGFR using a range of clinically relevant thresholds is
`shown in Figure 3. In the overall population, persistent
`
`580
`
`February 22, 2022
`
`Circulation. 2022;145:575–585. DOI: 10.1161/CIRCULATIONAHA.121.055459
`
`ORIGINAL RESEARCH
`
`ARTICLE
`
`Shaman et al
`
`Semaglutide and Liraglutide Outcomes
`
`Novo Nordisk Exhibit 2015
`Mylan Pharms. Inc. v. Novo Nordisk A/S
`IPR2023-00724
`Page 00006
`
`

`

`Figure 3. Effects of semaglutide and liraglutide versus placebo on time to the first persistent reduction in eGFR in the pooled
`population and subgroups according to eGFR at baseline.
`Effects of semaglutide and liraglutide versus placebo on time to the first persistent reduction in estimated glomerular filtration rate (eGFR)
`of 30%, 40%, 50%, and 57% from baseline in the pooled population and subgroups according to eGFR (mL/min/1.73 m2) at baseline. Time
`to persistent reduction of eGFR from baseline was analyzed independently from each other. Subgroup analyses were performed by including
`subgroup as a fixed factor and the interaction between subgroup and treatment. HR indicates hazard ratio.
`
`40% and 50% reductions in eGFR occurred less fre-
`quently in patients receiving semaglutide/liraglutide
`compared with placebo (HR, 0.86 [95% CI, 0.75–
`0.99]; P=0.039 and HR, 0.80 [95% CI, 0.66–0.97];
`P=0.023, respectively; Figure 3). The risk of reaching
`a persistent 30% eGFR reduction (HR, 0.92 [95%
`CI, 0.84–1.02]; P=0.10) or a persistent 57% reduc-
`tion (HR, 0.89 [95% CI, 0.69–1.13]; P=0.34) was not
`significantly reduced but showed similar directional re-
`sults (Figure 3).
`In subgroups of patients with an eGFR of 30 to
`<60 mL/min/1.73 m2 at baseline, persistent reduction
`in eGFR for all thresholds occurred in fewer patients in
`the semaglutide/liraglutide group compared with the
`placebo group, with a trend to greater effect sizes as
`eGFR thresholds increased. The HR (95% CI; percent-
`age semaglutide/liraglutide vs placebo) values were as
`follows: 30% reduction: 0.71 (0.59–0.85), P=0.0003
`(14.4% vs 19.2%); 40% reduction: 0.67 (0.52–0.86),
`P=0.0017 (7.7% vs 11.0%); 50% reduction: 0.56 (0.40–
`0.78), P=0.0006 (4.0% vs 6.8%); 57% reduction: 0.54
`
`(0.36–0.81), P=0.003 (2.6% vs 4.6%; Figure 3). The
`effect sizes appeared larger in subgroups of patients
`with eGFR 30 to <60 mL/min/1.73 m2 at baseline com-
`pared with other subgroups (eGFR ≥90, 60−<90, and
`<30 mL/min/1.73 m2 at baseline), especially for 30%
`and 57% reduction thresholds (Pinteraction=0.017 and
`0.035, respectively; Figure 3). On a continuous eGFR
`scale, the treatment effect associated with semaglutide/
`liraglutide versus placebo increased as baseline eGFR
`decreased (Figure S3).
`The results were broadly consistent for subgroups
`on the basis of baseline albuminuria (Figure 4), with
`separately statistically significant reductions in the risk
`of persistent 30%, 40%, and 50% reductions in eGFR
`with semaglutide/liraglutide compared with placebo
`in subgroups of patients with macroalbuminuria (HRs
`ranged from 0.78 [95% CI, 0.66–0.92]; P=0.004 for
`30% reduction in eGFR to 0.77 [95% CI, 0.60–1.00];
`P=0.050 for 50% reduction in eGFR) and microalbumin-
`uria or macroalbuminuria at baseline (HRs ranged from
`0.85 [95% CI, 0.76–0.97]; P=0.013 for 30% reduction
`
`Circulation. 2022;145:575–585. DOI: 10.1161/CIRCULATIONAHA.121.055459
`
`February 22, 2022
`
`581
`
`ORIGINAL RESEARCH
`
`ARTICLE
`
`Shaman et al
`
`Semaglutide and Liraglutide Outcomes
`
`Novo Nordisk Exhibit 2015
`Mylan Pharms. Inc. v. Novo Nordisk A/S
`IPR2023-00724
`Page 00007
`
`

`

`Figure 4. Effects of semaglutide and liraglutide versus placebo on time to the first persistent reduction in eGFR in the pooled
`population and subgroups defined by the level of albuminuria at baseline.
`Effects of semaglutide/liraglutide versus placebo on time to the first persistent reduction in estimated glomerular filtration rate (eGFR) of 30%,
`40%, 50%, and 57% from baseline in the pooled population and subgroups defined by the level of albuminuria at baseline. Time to persistent
`reduction of eGFR from baseline was analyzed independently from each other. Subgroup analyses were performed by including subgroup as a
`fixed factor and the interaction between subgroup and treatment. HR indicates hazard ratio.
`
`in eGFR to 0.76 [95% CI, 0.61–0.95]; P=0.016 for 50%
`reduction in eGFR).
`Similar patterns were observed in subgroups of
`patients stratified on the basis of a combination of
`baseline eGFR and albuminuria criteria (Figure S4).
`The effect of semaglutide/liraglutide therapy on per-
`sistent reduction in eGFR for all thresholds appeared
`to be larger in subgroups of patients with eGFR 30 to
`<60 mL/min/1.73 m2 and microalbuminuria or macro-
`albuminuria at baseline than in subgroups with baseline
`eGFR of ≥60 mL/min/1.73 m2 or normoalbuminuria as
`well as subgroups with eGFR <30 mL/min/1.73 m2 and
`microalbuminuria or macroalbuminuria at baseline.
`When persistent redu