`https://doi.org/10.1007/s11154-023-09807-3
`
`Effects of GLP-1 receptor agonists on neurological complications of
`diabetes
`
`Natalia García-Casares1,2,3,5 · Guillermo González-González1 · Carlos de la Cruz-Cosme2,4 · Francisco
`J Garzón-Maldonado2,4 · Carmen de Rojas-Leal1 · María J Ariza1,2,3 · Manuel Narváez1,2 · Miguel
`Ángel Barbancho1,2,3 · Juan Antonio García-Arnés1 · Francisco J. Tinahones1,2,4,6
`
`Accepted: 21 April 2023
`© The Author(s) 2023
`
`/ Published online: 26 May 2023
`
`Abstract
`Emerging evidence suggests that treatment with glucagon-like peptide-1 receptor agonists (GLP-1 RAs) could be an
`interesting treatment strategy to reduce neurological complications such as stroke, cognitive impairment, and peripheral
`neuropathy. We performed a systematic review to examine the evidence concerning the effects of GLP-1 RAs on neu-
`rological complications of diabetes. The databases used were Pubmed, Scopus and Cochrane. We selected clinical trials
`which analysed the effect of GLP-1 RAs on stroke, cognitive impairment, and peripheral neuropathy. We found a total of
`19 studies: 8 studies include stroke or major cardiovascular events, 7 involve cognitive impairment and 4 include periph-
`eral neuropathy. Semaglutide subcutaneous and dulaglutide reduced stroke cases. Liraglutide, albiglutide, oral semaglutide
`and efpeglenatide, were not shown to reduce the number of strokes but did reduce major cardiovascular events. Exenatide,
`dulaglutide and liraglutide improved general cognition but no significant effect on diabetic peripheral neuropathy has been
`reported with GLP-1 RAs. GLP-1 RAs are promising drugs that seem to be useful in the reduction of some neurological
`complications of diabetes. However, more studies are needed.
`
`Keywords Cognitive impairment · Stroke · Cardiovascular disease · Peripheral neuropathy · Cognitive impairment ·
`Alzheimer’s disease
`
` Natalia García-Casares
`Nagcasares@uma.es
` Francisco J. Tinahones
`fjtinahones@hotmail.com; fjtinahones@uma.es
`
`2
`
`1 Facultad de Medicina, Departamento de Medicina,
`Universidad de Málaga, Málaga, Spain
`Instituto de Investigación Biomédica de Málaga (I.B.I.M.A),
`Málaga, Spain
`3 Centro de Investigaciones Médico-Sanitarias (C.I.M.ES),
`Málaga, Spain
`4 Hospital Universitario Virgen de la Victoria de Málaga,
`Málaga, Spain
`5 Departamento de Medicina, Facultad de Medicina,
`Universidad de Málaga, Centro de Investigaciones Médico
`Sanitarias (C.I.M.E.S), Universidad de Málaga, Instituto
`de Investigación Biomédica de Málaga (IBIMA), Campus
`Universitario de Teatinos s/n., Málaga 29010, España
`6 Departamento de Medicina, Facultad de Medicina,
`Universidad de Málaga, Instituto de Investigación Biomédica
`de Málaga (IBIMA), Campus Universitario de Teatinos s/n.,
`Málaga 29010, España
`
`Abbreviations
`CVRF
` Cardiovascular risk factor.
`GLP-1
` RAs Glucagon-like peptide-1 receptor agonists.
`MACE
` Cardiovascular safety profile of currently avail-
`able diabetic drugs.
`
`1 Introduction
`
`People with diabetes have a high neurological risk, such
`as stroke, cognitive impairment, or diabetic neuropathy. In
`recent years we have witnessed the vertiginous progress in
`relation to new treatments for diabetes, which are not just
`limited to good glycaemic control, but can have beneficial
`effects on other organs such as the brain.
`Glucagon-like peptide-1 receptor agonists (GLP-1 RAs)
`currently play an important therapeutic role in the treatment
`of type 2 diabetes. They present similar properties to the
`human peptide as they reduce blood glucose and glycosylated
`haemoglobin levels, favouring insulin receptor sensitivity
`
`1 3
`
`Novo Nordisk Exhibit 2097
`Mylan Pharms. Inc. v. Novo Nordisk A/S
`IPR2023-00724
`Page 00001
`
`
`
`656
`
`Reviews in Endocrine and Metabolic Disorders (2023) 24:655–672
`
`and increasing insulin secretion. On the other hand, they can
`slow down gastric emptying, achieving an increase in sati-
`ety and a reduction in body weight. All the GLP-1-RAs (lix-
`isenatide, liraglutide, semaglutide, exenatide, dulaglutide,
`albiglutide, efpeglenatide) are administered via subcutane-
`ous (sc.) injection, except for semaglutide with both routes
`of administration, sc. and oral. Recent studies have shown
`that GLP-1 RAs also act on the nervous system, providing
`neuroprotective effects, mainly by controlling vascular risk
`factors and cardiovascular improvement [1].
`The cardiovascular safety of GLP-1-RAs in patients
`with type 2 diabetes has been demonstrated in multiple
`randomized controlled trials [2, 3]. The US Food and Drug
`Administration and the European Medicines Agency have
`required safety studies of these drugs on cardiovascular risk
`in patients with diabetes, which have allowed us to know
`their cardiovascular effects, including stroke, and further
`our understanding of its benefits in patients with type 2 dia-
`betes and high vascular risk.
`On the other hand, cognitive impairment is one more
`complication in persons with diabetes, presenting a cogni-
`tive profile with alterations mainly in the speed of informa-
`tion processing, verbal and visual memory, attention, and
`executive function [4]. Neuroimaging studies with MRI and
`brain PET have shown correlations with these neuropsycho-
`logical alterations and brain structural-functional alterations
`[5, 6]. Some studies suggest that there is an overlap between
`the pathophysiological mechanisms of Alzheimer’s disease
`and diabetes mediated by insulin resistance [7]. It is known
`that patients with Alzheimer’s disease present a decrease in
`insulin levels and an alteration of the signal in insulin recep-
`tors in the brain [7], favouring the deposition of amyloid
`beta and tau protein [8]. Experimental studies have shown
`the expression of GLP-1 receptors in areas such as the cere-
`bral hippocampus, especially in the dendrites of the pyrami-
`dal cells of the CA1 and CA3 region of this structure, which
`is crucial in the cognitive processes of learning and memory
`[9, 10]. Other studies have demonstrated the neuroprotec-
`tive role of GLP-1 analogues as they produce a reduction in
`oxidative stress and apoptosis, showing new brain connec-
`tions and neuroplasticity [11].
`Another complication of diabetes is peripheral neurop-
`athy affecting the peripheral nervous system. It can affect
`about 50% of patients with diabetes [12]. Symptoms of
`peripheral neuropathy present with varying degrees of
`numbness, tingling, or pain primarily in the distal parts of
`the extremities [13]. An early assessment of peripheral poly-
`neuropathy symptoms helps prevent ulcers, local infection,
`or sepsis, and even death [12]. Many risk factors for diabetic
`neuropathy have been identified, including the presence of
`cardiovascular risk factors [14] and studies have aimed to
`
`determine the possible benefit of GLP-1 RAs in the preven-
`tion or improvement of peripheral neuropathy.
`This study aimed to perform a systematic review to
`identify the evidence relating the effects of GLP-1 RAs on
`neurological complications of diabetes (stroke, cognitive
`impairment, and peripheral neuropathy).
`
`2 Methods
`
`2.1 Search strategy
`
`To undertake this systematic review, an exhaustive search
`was carried out through the Medline, Cochrane, and Sco-
`pus databases to identify those clinical trials related to the
`topic until January 2022. The keywords used were the same
`in the different databases: “GLP-1 Ras” [MeSH Terms]
`AND “stroke” [MeSH Terms] OR “cardiovascular disease”
`[MeSH Terms] OR “peripheral neuropathy” [MeSH Terms]
`OR “cognitive impairment” [MeSH Terms] OR “Alzheim-
`er’s disease” [MeSH Terms].
`Selection criteria.
`Studies were included in this systematic review if they
`met the following criteria: study design was a clinical trial or
`controlled clinical trial; participants were adults with a diag-
`nosis of either type 1 or 2 diabetes mellitus with GLP-1 RA
`treatment and a history of stroke or cardiovascular events,
`cognitive impairment, or peripheral neuropathy. Each study
`had to detail the number of physical events and mental
`points and then compare the results with the other interven-
`tion group. It was also necessary to specify if either stroke
`or a major cardiovascular event (MACE) was measured in
`the cardiovascular field, and what kind of tests and imag-
`ing were used in both cognitive impairment and peripheral
`neuropathy, so the starting point of the participants was also
`required. The exclusion criteria for this systematic review
`were clinical cases, reviews, meta-analyses, letters to the
`editor, single case studies, communications to conferences,
`other neurological diseases that were not addressed in the
`systematic review, animal studies and articles not written
`in English or Spanish. This systematic review is reported
`according to the Preferred Reported Items for Systematic
`reviews and Meta-Analyses (PRISMA) [15]. The PRISMA
`diagram illustrates the selection process of the studies and
`shows reasons for exclusion (Fig. 1).
`
`3 Results
`
`The search resulted in a total of 5954 articles, though after
`applying the inclusion and exclusion criteria, only 26 stud-
`ies were finally selected.
`
`1 3
`
`Novo Nordisk Exhibit 2097
`Mylan Pharms. Inc. v. Novo Nordisk A/S
`IPR2023-00724
`Page 00002
`
`
`
`Reviews in Endocrine and Metabolic Disorders (2023) 24:655–672
`
`657
`
`
`Fig. 1 Flowchart of the strategy
`search
`
`3.1 GLP-1 RAs and stroke/MACE
`
`This systematic review included 8 studies related to GLP-1
`RAs and stroke/MACE [16–23]. All were double blind, ran-
`domized, placebo-controlled trials, and all enrolled middle-
`or old-aged participants with a previous diagnosis of type 2
`diabetes as well as cardiovascular risk factors or cardiovas-
`cular disease. The study measure was specifically stroke in
`almost all articles [16–26], except for the AMPLITUDE-0
`trial [23] where it was MACE. Post-hoc analyses of these
`studies were included in the discussion (see Table 1).
`The primary objective of most of the large, double-blind,
`randomized, placebo-controlled trials that have evaluated
`the cardiovascular effects of GLP-1 RAs in patients with
`
`type 2 diabetes focused on evaluating the presence of a
`MACE, which was defined and generally included cardio-
`vascular death, non-fatal myocardial infarction and nonfatal
`stroke. The subgroup analyses of each article found the fol-
`lowing results in relation to the risk of stroke:
`In the ELIXA lixisenatide trial that evaluated patients
`with acute coronary syndrome (n = 6068 patients with an
`acute coronary event within 180 days before screening),
`subcutaneous (sc.) lixisenatide had no effect on the risk of
`nonfatal stroke over a median 25-month follow-up [16].
`In the EXSCEL cardiovascular event reduction study
`with exenatide, (n = 14,752 patients with or without estab-
`lished cardiovascular disease), extended-release exenatide
`
`1 3
`
`Novo Nordisk Exhibit 2097
`Mylan Pharms. Inc. v. Novo Nordisk A/S
`IPR2023-00724
`Page 00003
`
`
`
`658
`
`Reviews in Endocrine and Metabolic Disorders (2023) 24:655–672
`
`non-fatal stroke cases.
`of MACE events and
`suitable for the reduction
`Semaglutide sc. could be
`
`liraglutide.
`preventive effects of
`mediators of the MACE-
`tion, are thought to be
`extent UACR reduc-
`HbA1C, and to a lesser
`strokes.
`HF, but not the number of
`functional class I to III
`ous levels of NYHA
`diabetes despite previ-
`in patients with type 2
`the number of MACE
`Liraglutide sc. reduces
`cardiovascular events.
`diabetes and high risk of
`participants with type 2
`than MACE events in
`ing expanded-MACE
`more effective reduc-
`Liraglutide sc. is
`
`stroke cases.
`not for fatal or non-fatal
`of MACE events, but
`suitable for the reduction
`Liraglutide sc. could be
`stroke cases.
`nor for fatal or non-fatal
`tion of MACE events
`be suitable for the reduc-
`Lixisenatide sc. could not
`
`Conclusions
`
`(p = 0.04).
`non-fatal stroke cases compared to placebo
`
`STROKE-Semaglutide sc. reduced the number of
`
`(n = 1,649).
`versus placebo
`sc. (n = 1,648)
`semaglutide
`Once-weekly
`
`and at least one cardiovascular risk factor.
`aged > 60 years with diagnosis of diabetes
`kidney disease (stage 3 or higher), or
`disease, chronic heart failure or chronic
`diabetes with pre-existing cardiovascular
`Participants aged > 50 years with type 2
`
`N = 3,297
`
`trial. SUSTAIN-6.
`placebo-controlled
`double-blind,
`Randomized,
`
`(2016)
`al. [20]
`S.P. et
`Marso
`
`covariate.
`-UACR updated mean as time-dependent
`-UACR changed as time-dependent covariate.
`covariate.
`- HbA1C updated mean as time-dependent
`covariate.
`-HbA1C changed as time-dependent
`
`MACE
`
`(n = 4,672)
`or placebo
`(n = 4,668)
`liraglutide sc.
`Once-daily
`
`one cardiovascular risk factor.
`diagnosis of type 2 diabetes and at least
`vascular condition or aged > 60 years with
`of type 2 diabetes and at least one cardio-
`Participants aged > 50 years with diagnosis
`
`N = 9,340
`
`LEADER trial.
`placebo-controlled
`randomized,
`of double-blind,
`Post-hoc analysis
`
`(2020)
`al. [58]
`J.B. et
`Buse
`
`NYHA functional class I-III
`liraglutide sc. and placebo in patients with
`non-fatal stroke cases was observed between
`-No significant reduction in the number of
`patients with no heart failure
`strokes between liraglutide sc. and placebo in
`
`STROKE-No significant reduction in the number of
`
`compared with placebo.
`13.4% reduction in total expanded MACE
`relative risk reduction in total MACE and a
`-Liraglutide sc. was associated with a 15.7%
`pared to placebo (p = 0.3).
`the number of non-fatal stroke cases com-
`- Liraglutide sc. did not significantly reduce
`placebo (p = 0.16).
`the number of fatal stroke cases compared to
`STROKE-Liraglutide sc. did not significantly reduce
`
`MACE
`
`STROKE-Lixisenatide sc. did not significantly reduce
`measure
`Study
`
`Main findings
`
`placebo (p = 0.54).
`the number of stroke cases compared to
`
`(n = 4,672)
`or placebo
`(n = 4,668)
`liraglutide sc.
`Once-daily
`
`one cardiovascular risk factor.
`diagnosis of type 2 diabetes and at least
`vascular condition or aged > 60 years with
`of type 2 diabetes and at least one cardio-
`Participants aged > 50 years with diagnosis
`
`N = 9,340
`
`LEADER trial.
`placebo-controlled
`randomized,
`of double-blind,
`Post-hoc analysis
`
`(2020)
`[54]
`S. P.
`Marso
`
`(n = 4,672)
`or placebo
`(n = 4,668)
`liraglutide sc.
`Once-daily
`
`one cardiovascular risk factor.
`diagnosis of type 2 diabetes and at least
`vascular condition or aged > 60 years with
`of type 2 diabetes and at least one cardio-
`Participants aged > 50 years with diagnosis
`
`N = 9,340
`
`LEADER trial.
`placebo-controlled
`randomized,
`of double-blind,
`Post-hoc analysis
`
`(n = 4,672)
`or placebo
`(n = 4,668)
`liraglutide sc.
`Once-daily
`
`one cardiovascular risk factor.
`diagnosis of type 2 diabetes and at least
`vascular condition or aged > 60 years with
`of type 2 diabetes and at least one cardio-
`Participants aged > 50 years with diagnosis
`
`(n = 3,034)
`vs. placebo
`sc. (n = 3,034)
`Lixisenatide
`
`screening.
`coronary event within 180 days before
`of type 2 diabetes, who also had an acute
`Participants aged > 30 years with diagnosis
`
`N = 9,340
`
`N = 6,068
`
`trial; LEADER
`placebo-controlled
`double-blind,
`Randomized,
`trial; ELIXA
`placebo-controlled
`double-blind,
`randomized,
`Multicentre,
`
`Intervention
`
`Patients, nType of participants
`
`Study design
`
`(2019)
`[57]
`S. et al.
`Verma
`
`(2016)
`al. [19]
`S. P. et
`Marso
`
`(2015)
`al. [16]
`M.A. et
`Pfeffer
`[Ref.]
`Study
`
`Table 1 Glucagon-like peptide-1 receptor agonists and Stroke Mace/Events.
`
`1 3
`
`Novo Nordisk Exhibit 2097
`Mylan Pharms. Inc. v. Novo Nordisk A/S
`IPR2023-00724
`Page 00004
`
`
`
`Reviews in Endocrine and Metabolic Disorders (2023) 24:655–672
`
`659
`
`stroke cases.
`the number of non-fatal
`suitable for reducing
`Dulaglutide sc. may be
`fatal stroke cases.
`for reducing fatal or non-
`but might not be suitable
`the number of MACE
`Albiglutide sc. reduced
`non-fatal) cases.
`events or stroke (fatal or
`ing the number MACE
`be suitable for reduc-
`Exenatide sc. could not
`
`in placebo group (p = 0.017).
`glutide sc. group was significantly bigger than
`-Non-fatal stroke cases: the reduction in dula-
`than in placebo group (p = 0.34).
`tide sc. group was bigger, but non-significant,
`
`STROKE-Fatal stroke cases: the reduction in dulaglu-
`
`STROKE-Albiglutide sc. reduced the number of MACE
`
`does not reduce the incidence of strokes.
`-Subgroup analysis showed that albiglutide sc.
`(p = 0.0006).
`
`compared to placebo.
`the number of stroke cases (fatal or non-fatal)
`
`STROKE-Exenatide sc. did not significantly reduce
`
`disease presence.
`despite microvascular
`CV outcomes in patients
`glutide sc. reduce major
`Liraglutide sc. and sema-
`increased risk of MACE.
`was associated with an
`Microvascular disease
`
`their baseline BP.
`2 diabetes, irrespective of
`cial for patients with type
`glutide sc. may be benefi-
`Liraglutide sc. and sema-
`
`(p = 0.0136) and in SUSTAIN 6: (p = 0.0064).
`without microvascular disease in LEADER:
`cardiovascular events compared with those
`had an increased risk of major adverse
`-Participants with microvascular disease
`0.60–1.21).
`-BP stage 2 hypertension (HR: 0.85; 95% CI,
`0.37–1.03).
`-BP stage 1 hypertension (HR: 0.62; 95% CI,
`-BP elevated (HR: 0.43; 95% CI, 0.20–0.95).
`-BP normal (HR: 0.79; 95% CI, 0.40–1.56).
`Primary MACE in SUSTAIN-6:
`0.72–0.99).
`-BP stage 2 hypertension (HR: 0.84; 95% CI,
`0.60–0.90).
`-BP stage 1 hypertension (HR: 0.73; 95% CI,
`-BP elevated (HR: 1.21; 95% CI, 0.87–1.68).
`-BP normal (HR: 1.00; 95% CI, 0.75–1.32).
`Primary MACE in LEADER:
`
`Conclusions
`
`Main findings
`
`MACE
`
`MACE
`measure
`Study
`
`(n = 4,952).
`vs. placebo
`sc. (n = 4,949)
`Dulaglutide
`
`(n = 4,732).
`vs. placebo
`sc. (n = 4,731)
`Albiglutide
`
`(n = 7,396).
`vs. placebo
`(n = 7,356)
`Exenatide sc.
`trial.
`SUSTAIN 6
`(n = 1,649) in
`versus placebo
`sc. (n = 1,648)
`semaglutide
`Once-weekly
`LEADER trial.
`(n = 4,672) in
`or placebo
`(n = 4,668)
`liraglutide sc.
`Once-daily
`
`trial.
`SUSTAIN 6
`(n = 1,649) in
`versus placebo
`sc. (n = 1,648)
`semaglutide
`Once-weekly
`LEADER trial.
`(n = 4,672) in
`or placebo
`(n = 4,668)
`liraglutide sc.
`Once-daily
`
`Intervention
`
`
`
`Table 1 (continued)
`
`risk factors.
`cardiovascular events or cardiovascular
`sis of type 2 diabetes and either previous
`Participants aged > 50 years with diagno-
`
`N = 9,901
`
`disease.
`sis of type 2 diabetes and cardiovascular
`N = 10,793Participants aged > 40 years and diagno-
`
`lar events.
`of participants with previous cardiovascu-
`tes and the trial was designed to have 70%
`N = 14,752Participants with diagnosis of type 2 diabe-
`
`trial. REWIND
`placebo-controlled
`double-blind,
`randomized,
`Multicentre,
`
`HARMONY
`controlled trial.
`domized, placebo-
`Double-blind, ran-
`trial. EXSCEL
`placebo-controlled
`ized, double-blind,
`driven, random-
`Pragmatic, event-
`
`(2019)
`[21]
`H. et al.
`Gerstein
`(2018)
`[18]
`et al.
`dez A.F.
`Hernán-
`
`(2017)
`al. [17]
`R. R. et
`Holman
`
`N = 3,297
`TRIAL:
`TAIN-6
`-SUS-
`N = 9,340
`TRIAL:
`-LEADER
`
`SUSTAIN 6 trials.
`LEADER and
`placebo-controlled
`double-blind,
`of randomized,
`Post-hoc analysis
`
`(2020)
`[21]
`S. et al.
`Verma
`
`risk factor.
`of diabetes and at least one cardiovascular
`higher), or aged > 60 years with diagnosis
`ure or chronic kidney disease (stage 3 or
`cardiovascular disease, chronic heart fail-
`with type 2 diabetes with pre-existing
`SUSTAIN-6: Participants aged > 50 years
`risk factor.
`diabetes and at least one cardiovascular
`aged > 60 years with diagnosis of type 2
`at least one cardiovascular condition or
`with diagnosis of type 2 diabetes and
`LEADER: Participants aged > 50 years
`
`risk factor.
`of diabetes and at least one cardiovascular
`higher), or aged > 60 years with diagnosis
`ure or chronic kidney disease (stage 3 or
`cardiovascular disease, chronic heart fail-
`with type 2 diabetes with pre-existing
`SUSTAIN-6: Participants aged > 50 years
`risk factor.
`diabetes and at least one cardiovascular
`aged > 60 years with diagnosis of type 2
`at least one cardiovascular condition or
`with diagnosis of type 2 diabetes and
`LEADER: Participants aged > 50 years
`
`Patients, nType of participants
`
`Study design
`
`N = 3,297
`TRIAL:
`TAIN-6
`-SUS-
`N = 9,340
`TRIAL:
`-LEADER
`
`SUSTAIN 6 trials.
`LEADER and
`placebo-controlled
`double-blind,
`of randomized,
`Post-hoc analysis
`
`(2020)
`al. [22]
`L. A. et
`Leiter
`[Ref.]
`Study
`
`1 3
`
`Novo Nordisk Exhibit 2097
`Mylan Pharms. Inc. v. Novo Nordisk A/S
`IPR2023-00724
`Page 00005
`
`
`
`660
`
`Reviews in Endocrine and Metabolic Disorders (2023) 24:655–672
`
`the risk of MACE.
`Efpeglenatide sc. lowers
`
`not be proved.
`of semaglutide oral could
`Stroke-preventive effect
`MACE.
`expanded definition of a
`outcome that included an
`est for the composite
`reduction was great-
`CV risk. The absolute
`population at moderate
`or fatal outcomes in a
`the event burden of CV
`Dulaglutide sc. reduced
`affect stroke severity.
`2 diabetes but does not
`stroke in people with type
`cally relevant ischaemic
`use might reduce clini-
`Long-term dulaglutide sc.
`
`Conclusions
`
`
`
`ous
`BP: Blood Pressure. CV: Cardiovascular. GLP-1 RAs: Glucagon-like Peptide-1 Receptor Agonists. HBA1C: Haemoglobin A1c. MACE: Major Adverse Cardiovascular Events. SC: subcutane-
`
`(p = 0.007).
`treated with efpeglenatide sc. vs. placebo
`number of MACE events in those participants
`-There was a significant reduction in the
`
`MACE
`
`number of strokes.
`between semaglutide oral or placebo in the
`
`STROKE-There were no significant differences
`
`(n = 1,359).
`vs. placebo
`sc. (n = 2,717)
`Efpeglenatide
`(n = 1,592).
`or placebo
`oral (n = 1,591)
`semaglutide
`Once-daily
`
`CV risk factors.
`kidney disease and one or more additional
`CV disease, or ≥ 50 years old with chronic
`diagnosis of type 2 diabetes and previous
`
` N = 4,076Participants aged ≥ 18 years old with
`
`tes and cardiovascular risk factors.
`or aged > 60 years with diagnosis of diabe-
`vascular disease or chronic kidney disease,
`of type 2 diabetes and established cardio-
`
` N = 3,183Participants aged > 50 years with diagnosis
`
`AMPLITUDE-0
`trial.
`placebo-controlled
`Randomized,
`PIONEER-6
`controlled trial.
`domized, placebo-
`double-blind, ran-
`Event-driven,
`
`(2021)
`al. [23]
`H. C. et
`Gerstein
`
`(2019)
`[22]
`M. et al.
`Husain
`
`expanded MACE (p = 0.028).
`-Dulaglutide sc. reduced the incidence of total
`(p = 0.022).
`total MACE or non-cardiovascular deaths
`- Dulaglutide sc. reduced the incidence of
`rhagic stroke compared to placebo (p = 0·89).
`-Dulaglutide sc. had no effect on haemor-
`compared to placebo (p = 0·042).
`-Dulaglutide sc. reduced disabling-stroke
`compared to placebo (p = 0·012).
`
`STROKE-Dulaglutide sc. reduced ischaemic stroke
`measure
`Study
`
`Main findings
`
`(n = 4,952).
`vs. placebo
`sc. (n = 4,949)
`Dulaglutide
`
`Intervention
`
`MACE
`Expanded
`vs.
`MACE
`
`(n = 4,952).
`vs. placebo
`sc. (n = 4,949)
`Dulaglutide
`
`N = 12,133Participants aged ≥ 50 years with diagnosis
`
`of type 2 diabetes.
`
`N = 12,133Participants aged ≥ 50 years with diagnosis
`
`of type 2 diabetes.
`
`Patients, nType of participants
`
`Study design
`
`REWIND trial.
`placebo-controlled
`randomized,
`double-blind,
`of the multicentre,
`Post-hoc analysis
`REWIND trial.
`placebo-controlled
`randomized,
`double-blind,
`of the multicentre,
`Post-hoc analysis
`
`(2020)
`al. [60]
`G.R. et
`nais
`Dage-
`
`(2020)
`al. [59]
`H.C. et
`Gerstein
`[Ref.]
`Study
`
`Table 1 (continued)
`
`1 3
`
`Novo Nordisk Exhibit 2097
`Mylan Pharms. Inc. v. Novo Nordisk A/S
`IPR2023-00724
`Page 00006
`
`
`
`Reviews in Endocrine and Metabolic Disorders (2023) 24:655–672
`
`661
`
`also had no effect on the incidence of fatal or nonfatal stroke
`[17].
`In the HARMONY study (n = 9463 patients with estab-
`lished coronary heart disease, cerebrovascular disease, or
`peripheral vascular disease), albiglutide, despite lowering
`MACE, had no effect on the risk of fatal or nonfatal stroke
`in the subgroup analysis during a follow-up of 1.6 years
`[18]. This drug was later withdrawn in 2018 due to com-
`mercial problems.
`In the LEADER study with liraglutide, in the evaluation
`of cardiovascular events [n = 9340 patients, ≥ 50 years with
`established cardiovascular disease (coronary event, cere-
`brovascular event, or peripheral vascular disease), stage ≥ 3
`chronic kidney disease, or chronic heart failure according
`to New York Heart Association class II or III or ≥60 years
`with at least one cardiovascular risk factor (CVRF)], lira-
`glutide demonstrated a significant reduction in MACE by
`13% at the expense of a reduction in cardiovascular mortal-
`ity. However, subgroup analysis showed a non-significant
`trend in the effect on the incidence of fatal or nonfatal stroke
`or transient ischaemic attack during a follow-up of 3.8 years
`[19].
`The SUSTAIN-6 study [20], which evaluated the long-
`term cardiovascular effects of semaglutide sc., in subjects
`with type 2 diabetes and HbA1c ≥ 7%, treated with up to
`two oral hypoglycaemic agents, with or without basal or
`premixed insulin, [n = 3297 patients, ≥ 50 years with estab-
`lished cardiovascular disease (coronary event, cerebrovas-
`cular event, or peripheral vascular disease), chronic kidney
`disease stage ≥ 3, or New York Heart Association class II
`or III chronic heart failure or ≥ 60 years with at least one
`CVRF)], found a significant reduction in the primary end-
`point of MACE appearance by 26% (6.6% vs. 8.9%, HR
`0.74, 95% CI 0.58–0.95). The subgroup analysis showed
`that this reduction of the primary objective was achieved
`at the expense of the reduction in nonfatal stroke, since
`treatment with semaglutide sc. reduced the risk of nonfa-
`tal stroke by 39% (1.6% vs. 2.7%, HR 0.61, 95% CI 0.38–
`0.99; p = 0.04) during a follow-up of 2.1 years. On the other
`hand, at the beginning of the study, 76.3% of the patients
`(n = 2514) were receiving preventive treatment for stroke:
`85% of the patients were taking antiaggregants and 5.9%
`anticoagulants. Of the total of 3297 patients, during the
`study 71 patients presented a nonfatal stroke: 27 treated
`with semaglutide sc. [25 (1.5%) ischaemic and 2 (0.1%)
`haemorrhagic] vs. 44 with placebo [37 (2.2%) ischaemic,
`4 (0.2%) haemorrhagic, and 3 (0.2%) unclassifiable stroke].
`The REWIND study [21] with dulaglutide sc. [n = 9901
`patients, ≥ 50 years with established cardiovascular disease
`or ≥ 60 years with at least two CVRFs], met the primary
`efficacy endpoint, showing that weekly dulaglutide signifi-
`cantly reduced MACE (12.0% vs. 13.4%, HR 0.88, 95%
`
`CI 0.79–0.99; p = 0.026), with a significant 24% reduction
`in nonfatal stroke (3.2% vs. 4.1%, HR 0.76, 95% CI 0.61–
`0.95; p = 0.017) during a follow-up of 5.4 years.
`In the PIONEER-6 study with oral semaglutide [n = 3183
`patients, ≥ 50 years with established cardiovascular or renal
`disease, or ≥ 60 years with CVRFs], no significant efficacy
`of oral semaglutide could be demonstrated in the reduction
`in nonfatal stroke (HR 0.74, 95% CI 0.35–1.57) during a
`2.1-year follow-up [22].
`Finally, in AMPLITUDE-0 with efpeglenatide [n = 4076
`patients, ≥ 18 years with previous cardiovascular disease
`or ≥ 50 years with chronic kidney disease and 1 or more
`CVRFs], the efficacy objective was also achieved, demon-
`strating a significant decrease in MACE (7.0% vs. 9.2%,
`HR 0.73, 95% CI 0.58–0.92; p = 0.007). However, disaggre-
`gated stroke data are not yet available [23].
`
`3.2 GLP-1 RAs and cognitive impairment
`
`We found 7 articles related to GLP-1 RAs and cognitive
`impairment (see Table 2) [24–30], 5 of which were double
`blind, randomized, placebo-controlled trials [24, 25, 27,
`28, 30], while that of Zhang et al. [26] was a randomized
`clinical trial and that of Li et al. [29] a prospective, par-
`allel, open-label study. All 7 studies included middle- or
`old-age participants. The studies of Cukierman-Yaffe et al.
`[24], Zhang et al. [26], Li et al. [29] and Vadini et al. [30]
`included participants with a diagnosis of type 2 diabetes,
`whereas the studies of Watson et al. [25], Mullins et al. [27]
`and Gejl et al. [28] excluded these patients. However, only
`Zhang et al. [26], Mullins et al. [27] and Gejl et al. [28]
`included participants with mild cognitive impairment; the
`remaining 4 studies enrolled those with no cognitive out-
`comes at baseline.
`All studies measured glucose plasma biomarkers in
`blood test, except for Cukierman-Yaffe et al. [24] and Gejl
`et al. [28]. However, Mullins et al. [27] also measured CSF
`biomarkers as well. All 7 studies used different neuropsy-
`chological tests for cognitive assessment. Neuroimaging
`consisted of MRI acquisition before and after the interven-
`tion in all studies, except for the studies of Cukierman-Yaffe
`et al. [24] and Vadini et al. [30], in which no neuroimaging
`was performed.
`Cukierman-Yaffe et al. [24], using dulaglutide sc. vs.
`placebo, observed a non-significant reduction in cognitive
`impairment; p = 0.11. However, after post-hoc adjustments
`for individual standardized baseline scores, they found a
`significant reduction in the number of cognitive outcomes;
`p = 0.0018. In this sense, Zhang et al. [26], included type 2
`diabetes mellitus obese patients under GLP1RA treatment
`vs. type 2 diabetes mellitus non-obese patients under met-
`formin vs. control subjects, and found an increase in general
`
`1 3
`
`Novo Nordisk Exhibit 2097
`Mylan Pharms. Inc. v. Novo Nordisk A/S
`IPR2023-00724
`Page 00007
`
`
`
`662
`
`Reviews in Endocrine and Metabolic Disorders (2023) 24:655–672
`
`changes).
`detecting these
`not focused on
`its design was
`groups (although
`ences between
`cognitive differ-
`sc. showed no
`-Liraglutide
`
`-No cognitive differences between study groups.
`placebo.
`ity within the DMN in the active group relative to
`-Significant improvement in intrinsic connectiv-
`anterior medial frontal structures.
`connectivity between bilateral hippocampal and
`- Higher FPG was associated with decreased
`slightly for the active group (p = 0.06).
`-Glucose tolerance measured at 120 min declined
`
`MRI
`
`cose tolerance
`-Impaired glu-
`glucose,
`plasma
`-Fasting
`biomarkers:
`Plasma
`
`Matrix Reasoning
`WASI: Vocabulary
`WAIS-III
`SDFR
`RCFT
`Pegboard
`D-KEFS
`CVLT-II
`BVRT
`BNT
`
`(n = 16).
`vs. placebo
`sc. (n = 25)
`Liraglutide
`
`of education.
`least 12 years
`of > 27 and at
`MMSE score
`2 diabetes,
`absence of type
`years old,
`between 45–70
`Participants
`
`N = 43
`
`cognition.
`groups in general
`ferences between
`did not show dif-
`Liraglutide sc.
`
`Conclusions
`
`-No apoE subgroups reported.
`-No differences in aβ deposition and cognition.
`pared with placebo, which underwent decline.
`-Liraglutide sc. group, preserved CMRglu com-
`
`PET)
`([18 F] FDG
`CMRglu and
`PIB PET),
`CBF ([11 C]
`deposition and
`Aβ-protein
`
`NO
`
`examination)
`(brief cognitive
`WMS-IV
`
`(n = 20)
`vs. placebo
`sc. (n = 18)
`Liraglutide
`
`2 diabetes
`nosed with type
`and not diag-
`er’s Disease
`with Alzheim-
`Participants
`
`N = 38
`
`NeuroimagingMain findings
`
`CSF tests
`Blood and
`
`cal and other tests
`InterventionNeuropsychologi-
`
`participants
`Type of
`
`Patients,
`
`N
`
`design
`Study
`
`Table 2 Glucagon-like peptide-1 receptor agonists and Cognitive Impairment
`
`trial.
`trolled
`con-
`cebo-
`pla-
`ized,
`dom-
`ran-
`blind,
`ble-
`Dou-
`trial.
`clinical
`trolled,
`con-
`cebo-
`pla-
`ized
`dom-
`ran-
`blind,
`ble-
`Dou-
`
`(2018)
`[25]
`K.T. et al.
`Watson
`
`(2017)
`al. [28]
`Gejl M. et
`publication
`year of
`[Ref.] and
`Study
`
`1 3
`
`Novo Nordisk Exhibit 2097
`Mylan Pharms. Inc. v. Novo Nordisk A/S
`IPR2023-00724
`Page 00008
`
`
`
`Reviews in Endocrine and Metabolic Disorders (2023) 24:655–672
`
`663
`
`memory.
`in attention and
`and improvement
`enriched EVs
`neuronal-origin
`in Aβ42 in
`of a decrease
`sible exception
`with the pos-
`good findings
`not demonstrate
`exenatide sc. can-
`outcomes with
`AD-related
`
`Conclusions
`
`
`
`plexity, gyrification, or sulcal depth.
`GM volume or cortical thickness, cortical com-
`-No significant effect of exenatide on preserving
`in MRS metabolites.
`-No significant changes were observed over time
`group (p = 0.045).
`which decreased over time in exenatide-treated
`CSF, plasma and EV, except from Aβ42 in EVs,
`-No significant treatment effects on biomarkers in
`maximum digit-span forward (p = 0.17).
`Digit-span forward total score (p = 0.006) and
`-Better performance for exenatide sc. group in
`
`MRS
`MRI
`
`IRS-1.
`TAU and
`lated proteins
`-Phosphory-
`-Aβ42
`-Aβ40
`biomarkers:
`EV
`-Homa2
`-OGTT
`insulin
`-Fasting
`glucose
`-Fasting
`-Aβ42
`-Aβ40
`biomarkers:
`Plasma
`-p181-tau
`-Total TAU
`-Aβ42
`biomarkers:
`CSF
`
`Memory Subtest.
`Scale Logical
`Wechsler Memory
`Intelligence Scale,
`Wechsler Adult
`UPSIT,
`Parts A & B,
`Trail-making Test
`MMSE,
`FAB,
`Subtests,
`ward & Backward
`Digit Span For-
`Digit-Symbol and
`CVLT,
`Clock drawing,
`CDR-sob,
`CDR,
`BVRT,
`Test,
`Boston Naming
`ANART,
`ADCS-ADL,
`ADAS-Cog,
`
`(n = 10).
`vs.