ExP.erf
`Opinion
`
`Introduction
`1.
`2. Exenatide
`3. GLP-1 analogs under
`investigation
`4. DPP-4 inhibitors
`5. Safety of DPP-4 inhibitors
`6. Exenatide versus DPP-4
`inhibitors
`7. Expert opinion
`
`informa
`
`healthcare
`
`Review
`
`lncretin mimetics and dipeptidyl
`peptidase 4 inhibitors in clinical
`trials for the treatment of
`type 2 diabetes
`
`N asser M ikhail
`Olive-View-UCLA Medical Center, Endocrinology Division, Department of Medicine, Sylmar, C4, USA
`
`Background: Exenatide is an incretin mimetic, while sitagliptin and vildagliptin
`are
`incretin enhancers used as adjunctive
`therapy
`in patients with
`type 2 diabetes failing oral agents. Sitagliptin and vildagliptin can also be
`used as monotherapy in patients with type 2 diabetes uncontrolled by diet.
`Objective: To provide a critical review of clinical trials of exenatide, sitagliptin
`and vildagliptin . Method: Review of Phase Ill clinical trials based on Medline
`search published up to April 2008. Results: The use of exenatide is associated
`with reduction in average hemoglobin A1c (HbA1c) levels of approximately
`0.8% compared with baseline. The corresponding reduction with either
`sitagliptin or vildagliptin is 0.7%. The actions of incretin-based drugs predomi(cid:173)
`nantly target postprandial hyperglycemia. Treatment-related hypoglycemia
`is generally mild, and mainly occurs when used with sulfonylureas (SUs) .
`Exenatide treatment leads to a mild weight loss of around 2 kg after
`30 weeks, whereas sitagliptin and vildagliptin have generally neutral effect
`on weight. Sitagliptin and vildagliptin are well tolerated in trials lasting up
`to 52 weeks. Meanwhile, 5 - 10% of patients cannot tolerate exenatide due
`to adverse effects, mainly nausea and vomiting . The three drugs are limited
`by the lack of long-term safety and efficacy data, as well as by their high
`cost. Conclusion: Exenatide, sitagliptin and vildagliptin are useful add-on
`therapy for type 2 diabetes that is suboptimally controlled on oral agent s,
`particularly when there is concern about weight gain and hypoglycemia, or
`when postprandial hyperglycemia is the major cause of inadequate glycemic
`control. Sitagliptin and vildagliptin may be used as monotherapy in patients
`who cannot tolerate metformin or SU, and sitagliptin may be used as
`alternative to metformin in renal insufficiency.
`
`Keywords: dipeptidyl peptidase inhibitors, exenatide, increti ns, sitagli ptin ,
`rype 2 d iabetes, vi ldagl iptin
`
`Expert Opin. lnvestig. Drugs (2008) 17(6):845-853
`
`1. Introduction
`
`Glucagon-like peptide- 1 (G LP- 1) and glucose-dependent insul inotropic polypeptide,
`also called gastric inh ibi tory polypeptide (GIP), are the two mai n physiological
`incretins synthesized in the intestinal tract [1]. Al though both incretins enhance
`glucose-induced insuli n secretio n after meals, research has focused on GLP- 1 as a
`candidate anti-diabetic agent fo r several reasons. First, it is estimated that GLP-1
`accounts for at least 50% of the total incretin activity [2,3]. Second, on a molar
`basis, the effect of exogenous GLP-1 o n insulin secretio n in hea lthy subj ects
`is substantially greater than that of G IP [4 -6]. Thi rd, in additio n to its ins ul i(cid:173)
`notropic action, GLP- 1, but not G IP, inhib its glu cago n release, delays gastric
`emptying, and may promote early satiety
`[4,6,7-9]. Fourth, in patients with
`
`10 15 1 7/1354ft?j§q{\~Uf?alWa~6Q8~forma UK Ltd ISSN 1354-3784
`at the NLM and may b<e·
`Sub<ject US Copyright Laws
`
`845
`
`Mylan EX 1003, Page 1
`
`

`
`lncretin mimetics and dipeptidyl peptidase 4 inhibitors in clinical trials for the treatment of type 2 diabetes
`
`type 2 diabetes, GLP-1 administered in physiological [4,7]
`and supraphysiological [7,10,11] doses proved a potent insulin
`secretagogue, whereas GIP given in approximate equimolar
`doses had minimal [7, 12], or no [5] effect on insu lin secretion.
`Finally, the response of native GLP- 1 to meals in type 2
`diabetes is decreased [13,1 4] or absent [1 5], a defect that may
`exacerbate postprandial hyperglycemia.
`Despite the beneficial actions of GLP- 1 and GIP on glucose
`control, their use as anti-diabetic agents was impractical due
`to their short half-lifes as resu lt of their rapid inactivation by
`a protease called dipeptidyl peptidase type 4 (DPP-4) [I].
`For examp le, the half-life of GLP- 1 is around 2 min after
`[IG] . Thus, two approaches have been
`i.v. administration
`undertaken to overcome this problem. The first consists in
`the development of GLP- 1 ana logs, also called incretin
`mimetics, that bind to the GLP-1 receptors with the same
`affinity as GLP- 1 bm resist the degradation by DPP-4. The
`second is to design drugs that inhibit the action of DPP-4,
`ca lled incretin enhancers. The latter agents prolong the
`effects of native GLP- 1 and GIP and increase their serum
`levels approximately twofold. This article will focus on the
`incretin mimetic exenatide (Byena), which was approved by
`the US FDA in April 2005 and became available in the
`UK in May 2007, and the two incretin enhancers sitagliptin
`Oanuvia) and vildagliptin (Galvus). Sitagliptin received FDA
`approval in October 2006. Vildagliptin has not yet been
`approved in the USA, bm the drug is used in several other
`countries such as Mexico and Brazil.
`
`2. Exenatide
`
`2.1 Clinical efficacy
`2.7.7 Exenatide in combination with oral
`anti-diabetic therapy
`Exenatide was not eva luated as monotherapy in clinical trials.
`However, the drug was assessed as adjunctive therapy in three
`trials of similar design, including > 1400 obese patients with
`type 2 diabetes uncontrolled with metformin [1 7], sulfonylurea
`(SU) [1 8], or both [19]. After 30 weeks, average reductions in
`HbA1 c levels with the high dose of exenatide (1 0 pg b.i.d.)
`were approxim ately 0.8 and 1.0% compared with baseline
`and placebo, respectively. Similar reductions in HbA1c values
`were reported in a smaller trial (n = 232) of shorter duration
`(16 weeks), in which exenatide was evaluated as add-on therapy
`in patients with type 2 diabetes suboptimally controlled on a
`thiazolidenedione (TZD) and metformin [20]. A long-acting
`release (LAR) formulation of exenatide with a median half-life
`of 2 weeks is under investigation for once-weekly injection.
`In a Phase II, randomized, double-blind trial, 2 mg exenatide
`LAR given s.c. by the investigators to patients with type 2
`diabetes on metformin therapy red uced HbA1c values by
`an average of 1.7% after 15 weeks (n = 15), compared
`with 0.4% increase with placebo (n = 14) [21].
`At the end of the previous five trials, the average proportions
`of subjects who achieved 1-IbA! c value of ~ 7.0%, the
`
`optimum level recommended by the American Diabetes
`Association [22), were 45 and 10% in the exenatide and
`placebo groups, respectively [23]. Clearly, besides drug efficacy,
`these proportions depend on the baseline mean levels of
`1-IbAlc, which were in the range 7.9 - 8.7% [17-20] .
`
`2.1.2 Exenatide compared with insulin
`Exenatide was compared with insulin glargine in 549 patients
`with type 2 diabetes (mean baseline HbAl c of around 8.3%)
`on a background therapy that consists ofSU plus metformin [24].
`After 26 weeks,
`the mean reduction
`in HbAl c
`levels
`was 1.1 o/o in both groups. In another trial, exenatide was
`compared with biphasic insulin asparr (formed of 30%
`short-acting insulin aspart and 70%
`intermediate-acting
`insulin) as adjunctive
`therapy
`in patients with
`type 2
`diabetes (n = 50 I) inadequately controlled on metformin
`plus SU (mean baseline HbA1c 8.6%) [25]. After 52 weeks,
`no significant differences
`in 1-IbAlc
`reductions were
`found between the exenatide and biphasic insulin asparr
`groups: I and 0.9%, respectively [25].
`In the previous two trials [24 ,25], the mean daily doses of
`insulin glargine and biphasic insulin aspart at the srudy ends
`were 26 and 24 units, respectively, suggesting that exenatide
`efficacy (10 pg b.i.d.) may be equivalent to mean daily
`insulin doses close to that range. However, more studies are
`needed to examine
`the benefits and risks of switching
`from insulin to exenatide therapy. Unti l these studi es become
`available, such a strategy is not recommended, particularly
`in patients whose diabetes is not controlled on relatively
`high doses of insulin. For
`instance,
`in an exploratory
`study of 49 subjects with type 2 diabetes having mean
`baseline HbAlc values of approximately 8.1 o/o while receiving
`insulin doses > 40 units/d, the substitution of exenatide
`for insulin resulted in further deterioration of glycemic
`control in 40% of patients, and lack of improvement in the
`remaining 60% of patients [26] .
`
`.
`111
`
`2.1.3 Effect of exenatide on body weight
`One advantage of exenatide
`is
`that
`improvement
`glycem ic control is generally associated with mild weight
`loss. T he average weight Joss with the highest dose of
`exenatide was in the range 1.6 _ 2.8 kg compared with
`baseline, and 0.7 - 2.5 kg compared with placebo [17-19].
`When compared with insulin glargine, patients randomized
`to exenatide lost an average weight of 2.3 kg, as opposed to
`a weight gain of 1.8 kg in the glargine group [24]. Similarly,
`patients achieved a mean weight loss of2.5 kg with exenatide
`therapy, compared with a weight gain of 2.9 kg with
`biphasic insulin aspart [25].
`The reasons of weight loss associated with exenatide
`treatment are not well defined. Decreased appetite, sensation
`of early fullness, nausea, and the delay in gastric emptying
`may be contributing factors. Open-label extension studies of
`a selected cohort of 3 14 patients showed that weight loss
`with exenatide might be progressive, reaching an average
`
`846
`
`Expert Opin. lnvestig. Drugs (2008) 17(6)
`Th is mat~r i a l w as~o,p i e.d
`at the NLM and may ~e,
`Su~jec.t USCopyr igllt Law s
`
`Mylan EX 1003, Page 2
`
`

`
`Mikhail
`
`[27). Importantly, this weight
`of 4.4 kg after 82 weeks
`reduction was associated with significant improvement in
`triglycerides,
`cardiovascular risk factors such as plasma
`high-density
`lipoprotein cholesterol
`levels and diastolic
`blood pressure [27].
`Contrary to the time course of the reduction in HbA1c
`levels, which generally plateau after 12 -
`16 weeks of
`exenatide therapy, the weight loss continues
`to progress
`after 16 weeks, albeit at a slower rate [25].
`
`2.2 Safety profile of exenatide
`2.2.1 Nausea
`Nausea was the commonest adverse effect of exenatide,
`reported by 45 - 51% of patients who used the drug
`(vs 7
`-
`23% with placebo),
`the highest
`frequency
`being reported during the first 8 weeks of therapy [1 7- 19,24·25] .
`It was generally mild-to-moderate
`in
`intensity and
`dose-related
`[1 7- 19). However,
`around 4% of patients
`withdrew from clinical trials due to severe nausea
`[17-1 9] .
`The etiology of nausea
`is not fully clear, but may be
`partly rel ated to the delay in gasrric emptying. Nausea
`did not seem
`to be
`the predominant factor
`in
`the
`weight loss induced by exenatide, because there was no
`significant correlation between change in body weight and
`the duration of nausea [1 7,19).
`
`2.2.2 Hypoglycemia
`Consistent with the glucose-dependent insulinotropic effect
`of exenatide, hypoglycemia caused by the drug was generally
`uncommon and mild-to-moderate in severity. Studies in healthy
`that glucagon and other hormonal
`volunteers suggest
`counterregulatory
`responses
`to
`hypoglycemia were
`preserved with short-term administration of exenatide [28].
`In clinical trials using metformin alone as background
`rreatmenr, the frequency of hypoglycemia in the exenatide
`and placebo groups was similar [1 7). However, hypoglycemia
`was more frequenr with exenatide compared with placebo
`in trials that included a SU as background therapy [1 8,19).
`Interestingly, it has been shown that GLP-1 and the SU
`glyburide exhibited a synergistic effect on insulin release
`in perfused rat pancreas and in a small group of eight
`[29) . Unexpectedly,
`obese patients with type 2 diabetes
`the incidence and severity of hypoglycemia with enexatide
`treatment were similar when compared with
`insulin
`glargine [24 ] and biphasic insulin aspart [25). This may be
`due in part due to the moderate insulin doses used in the
`previous two studies [24 ,25).
`
`2.2.3 Other adverse effects of exenatide
`Other adverse effects
`reported by higher proportions
`of exenatide-treated patients compared with placebo
`included diarrhea (9 -
`17% vs 4 - 8% with placebo),
`feeling jittery (12 - 15% vs 2 - 7% with placebo), dizziness
`(9 - 15% vs 6 - 7% with placebo) , constipation (9 vs 3%
`with placebo), sweating (8 vs 1% with placebo), and
`
`backache (6 vs 3% with placebo) [17- 19]. In the postmarketing
`period, 30 cases of pancreatitis possibly caused by exenatide
`were reported from the date of the drugs approval through
`to 31 December 2006 [30) .
`
`3. GlP-1 analogs under investigation
`
`Liraglutide (NN22 11) is another G LP- 1 analog with a lo ng
`duration of action (half-li fe of around 12 h) owing to its
`stability against DPP-4, albumin-binding acylated side chain, <md
`self-association, resulting in slow absorption &om subcutaneous
`tissue [31,32). It is given by a single daily s.c. injection [31),
`and is currently under evaluation in Phase III trials. In the
`largest randomized trial (n = 190) of liraglmide published to
`date, the mean reduction in HbAlc after 12 weeks was
`0.7% compared with placebo, which was similar to the
`reduction achieved by submaximal doses of glimepiride
`(mean daily dose 2.7 mg). The drug's main adverse effect
`was nausea, and body weight remained unchanged [32).
`
`4. DPP-4 inhibitors
`
`4.1 DPP-4 inhibitors as monotherapy
`several
`in
`Sitagliptin was evaluated as monotherapy
`double-blind placebo-controlled trials in patients with type 2
`diabetes [33. 34]. Given as a single daily oral dose of 100 mg,
`treatment with sitagliptin was associated with average
`reduction of HbAl c values of around 0.7%, compared with
`baseline and placebo values [33,34). The average proportio ns
`of patients who achieved HbAl c levels < 7% at the trial
`ends were 44 and 18% in th e sitagliptin and placebo groups,
`respectively [2 3). Similar decreases in HbAI c values were
`generally observed in trials using vildagliptin
`[35,36) . The
`latter was given as a 50 mg tablet b.i.d. or I 00 mg once
`daily, with no significant difference in effi cacy between the
`two dosing regimens [36] .
`
`4.2 DPP-4 inhibitors as part of combination therapy
`4.2. 7 DPP-4 inhibitors in addition to metformin
`When sitagliptin was used as add-on agent to ongoing
`rnetformin therapy, the mean reduction in HbA1 c values fro m
`a baseline of 8.0% was around 0.7% at 24 weeks, with no
`change in the placebo group [37) . Similarly, the addition of
`vildagliptin 50 mg b.i.d. to ongoing metformin therapy was
`associated with a mean decrease in HbA I c of 0.9 and 1.1 %
`compared with baseline (8.4%) and placebo, respectively [38].
`In a third trial form ed of six groups of patients, sitagliptin
`was evaluated as monotherapy as well as part of initial
`combination therapy with metformin. At 24 weeks, the
`mean decreases in HbAlc values from a mean baseline of
`8.8% were 0.7, 0.8, 1.1 , 1.4, 1.9, and a slight increase
`of 0.2% in the groups randomized to sitagliptin I 00 mg q.d .,
`metformin 500 mg b.i.d., metfonnin 1000 mg b.i.d. ,
`sitaglipcin 50 mg b.i.d . + metformin 500 mg b.i.d., sitagliptin
`50 mg b.i.d. + metformin 1000 mg b.i.d. , and placebo,
`
`Th &-'AAffJff.i'i\.!'il~et$1 .fJ.dugs (2008) 17(6)
`at t he NLM and m ay IJ.e
`Su l>j ect US Copyright Law s
`
`847
`
`Mylan EX 1003, Page 3
`
`

`
`lncretin mimetics and dipeptidyl peptidase 4 inhibitors in clinical trials for the treatment of type 2 diabetes
`
`respectively [39]. In the latter study, the effects of metformin
`and sitagliptin on HbA1c reduction were additive. Indeed,
`metformin was shown to increase GLP-1 plasma levels (40],
`possibly by acting as a weak DPP-4 inhibitor [4 1].
`
`4.2.2 DPP-4 inhibitors in addition to sulfonylurea
`In a randomized, double-blind trial, Hermansen et al.
`(42]
`evaluated the addi tional effect of sitagliptin versus placebo in
`44 1 patients with uncontrolled type 2 diabetes (mean baseline
`HbA1c value was 8.3%) on glimepiride alone or on
`glimepiride plus metformin. After 24 weeks, the addition
`of sitagliptin
`to glimepiride resulted
`in mean HbA1c
`reductions of 0.3 and 0.6% compared with baseline and
`placebo, respectively. In the subgroup of patients randomized
`to
`triple
`therapy formed of metformin, glimepiride and
`sitagliptin, the corresponding reductions in HbA1c values
`were 0.6 and 0.9% [42].
`Likewise, in patients with type 2 diabetes uncontrolled on
`glimepiride (mean HbA1c value of 8.5%), the addition of
`vildagliptin 50 mg once daily (n = 170) and 50 mg twice
`daily (n = 169) resulted in similar reductions in HbA1c
`values by around 0.6% after 24 weeks, compared with the
`addition of placebo (n = 176) [43]. The results of the previous
`two trials suggest that DPP-4 inhibitors may have some
`value in improving glycemic control when used in conjunction
`with a SU such as glimepiride, despite the fact that both
`classes of anti-diabetic agents stimulate insulin secretion.
`This improvement may result fi·om the glucagon-suppressive
`effect of DPP-4 inhibitors, and perhaps by boosting the
`insulin secretion after meals.
`action of glimep iride on
`Consistent with the latter hypothesis was the finding that the
`effects of sitagliptin and vildagliptin were more pronounced
`in decreasing postprandial glucose, whereas the reduction
`in fasting plasma glucose was minimal and did not reach
`statistical sign ificance (42,43].
`
`4.2.3 DPP-4 inhibitors in addition to thiazolidinediones
`The addition of sitagliptin (1 00 mg q.d.) to ongoing piogli(cid:173)
`tazone therapy (30 or 45 mg q.d.) for 24 weeks was associated
`with a mean reduction in HbA1c values of 0.8 and 0.7%
`compared with baseline (8.0%) and placebo, respectively [44].
`Similar results were obtained with the add ition of vildagliptin
`to patients with inadequately controlled type 2 diabetes
`(average baseline HbA1c 8.7%) on maximal pioglitazone
`doses (45 mg/d). Thus, the mean reductions in HbA1c values
`with vildagliptin 50 mg b.i.d. were 1.0 and 0. 7% compared
`with baseline and placebo, respectively [4 5].
`In a third trial, 607 drug-na"ive patients with uncontrolled
`type 2 diabetes (mean baseline HbA1c of around 8.7%)
`were randomized
`to four groups to receive pioglitazone
`30 mg q.d., vildagliptin 50 mg q.d. + pioglitazone 15 mg q.d.,
`vildagliptin 100 mg q.d. plus pioglitazone 30 mg q.d.,
`and vi ldagliptin 100 mg q.d. [46]. After 24 weeks, mean
`reductions in HbA1c from baseline were 1.4, 1.7, 1.9, and
`1.1% in the pioglitazone monotherapy, 50/1 5 mg combination,
`
`the vildagliptin mono(cid:173)
`100/30 mg combination, and
`therapy groups, respectively
`[46]. Thus, contrary to the
`combination of metformin plus sitagliptin [39], the efficacy
`of the pioglitazone/vildagliptin combination appears to be
`less than additive.
`
`4.2.4 DPP-4 inhibitors in conjunction with insulin
`In one trial , the addition of vildagliptin (50 mg b.i.d.) to
`ongoing insulin therapy in patients with advanced type 2 d ia(cid:173)
`betes modestly reduced HbA1c values by 0.5 and 0.3% after
`24 weeks compared with baseline and placebo, respectively [47].
`However, for unclear reasons, this reduction in HbA1c values
`was limited to the subgroup of patients aged > 65 years [47].
`
`4.3 DPP-4 inhibitors in direct comparison with
`existing anti-diabetic oral agents
`4.3.1 Comparison of DPP-4 inhibitors with metformin
`In a noninferiority trial, vildagliptin (50 mg b.i.d.) was
`(1 000 mg b.i.d.)
`compared with metformin
`in 780
`drug-na·ive patients. After 52 weeks, the average reductions
`in HbA1c values from baseline were sign ificantly greater
`with metformin compared with vildaglitin: 1.4 and 1.0%,
`respectively [48]. In another trial of 24-week duration, the
`placebo-subtracted
`reductions
`in HbA1c values with
`sitagliptin (1 00 mg once daily), metfonnin (500 mg b.i.d.)
`and metformin (1000 mg b.i.d.) were 0.8, 1.0, and 1.3%,
`respectively [39]. However, levels of statistical significance were
`not reported. In the previous two trials, patients on metformin
`lost an average weight of 1.6 - 1.9 kg, but gastrointestinal
`adverse effects were more frequently reported [39,48] .
`
`4.3.2 Comparison of DPP-4 inhibitors with
`sulfonylureas
`In a noninferiority trial, sitagliptin was compared with glipizide
`as add-on therapy in > 1000 patients with inadequate
`glycemic control on metformin [49]. After 52 weeks, both
`groups had sim ilar reductions in HbA1c values of approxi(cid:173)
`mately 0.7% versus baseline. However, the mean daily dose
`of glipizide was submaximal (around 10 mg), and withdrawal
`rates due to lack of efficacy were higher with sitagliptin
`compared with glipizide: 86 of 588 patients (15%) versus
`58 of 584 (1 0%) patients [IJ9]. On the other hand, sitagliptin
`was associated with lower rates of hypoglycemia (5 vs 32%
`of patients), and a sl ight weight loss of 1.5 kg compared
`with 1.1 kg of weight gain with glipizide [49].
`
`4.3.3 Comparison of DPP-4 inhibitors with
`thiazolidinediones
`In drug-na·ive patients with type 2 diabetes, vildagl iptin
`(50 mg b.i.d.) and rosiglitazone (8 mg once daily) decreased
`HbA1 c values by 1.1 and 1.3%, respectively, after 24 weeks,
`meeting the statistical criterion of noninferiority of vildagliptin
`relative to rosiglitazone. Patients on rosiglitazone had an
`average weight gain of 1.6 kg, while vildagliptin had no
`effect on weight [50].
`
`848
`
`Expert Opin. ln vestia .. Druas (2008) 17(6)
`Th1s m atlfn al vrascoJ)Ie.d
`at th e NLM and m ay ~e
`Subj ect US Copyright Law s
`
`Mylan EX 1003, Page 4
`
`

`
`Mikhail
`
`In another trial including patients with type 2 diabetes
`inadequately controlled on metformin (mean HbA1 c 8.4%),
`additional treatment with vildagliptin (50 mg b.i.d.) was
`compared with pioglitazone given in submaximal doses
`(30 mg/day)
`[5 1]. After 24 weeks,
`the
`reductions
`in
`mean HbA1c values were similar in the vildagliptin and
`pioglitazone groups: 0.9 and 1.0%, respectively. Mean
`weight ga in was significantly greater in the pioglitazone
`group compared with the vildagliptin group - 1.9 and 0.3 kg,
`respective! y [5 I].
`
`4.3.4 Comparison of DPP-4 inhibitors with acarbose
`In a single
`randomized double-blind
`trial, vildagliptin
`(50 mg b.i.d.) was compared with the a-glucosidase inhibitor
`acarbose (up to l 00 mg t.i.d.) in drug-na'ive patients with
`type 2 diabetes [52]. After 24 weeks, the reductions in HbA1 c
`levels from a baseline of 8.6% were similar in the vildagliptin
`group (n = 441) and the acarbose group (n = 220), 1.4 and
`1.3%, respectively. The relatively high reductions in HbA1c
`values with either vildagliptin or acarbose observed in this
`trial may be attributed to the short duration of diabetes
`(mean duration since diagnosis was 1.2 years), and the
`additional benefit of lifestyle changes [52]. Gastrointestinal
`adverse effects were reported by 25% of acarbose-rreated
`patients, compared with 12% of patients receiving vildagliptin.
`Body weight significantly decreased with acarbose therapy
`(1.7 kg) and marginally decreased
`in
`the vildagliptin
`group (0.4 kg) [52).
`Taken together, the available limited head-to-head trials
`suggest that current DPP-4 inhibitors are less effective in
`lowering HbA1c values compared with metformin [39.18),
`and probably slightly less effective
`than maximal doses
`of SUs [19] and pioglitazone [46,51]; but they appear to be
`similarly effective to rosiglitazone
`[50] and acarbose
`[52].
`Meanwhile, the effects on DPP-4 inhibitors on variables
`other than HbAlc levels - such as body weight, hypoglycemia,
`and profile of adverse effects - are different. These variables
`must be considered when selecting the most appropriate
`patients to receive the DPP-4 inhibitors, as outlined in
`the last section.
`
`s. Safety of DPP-4 inhibitors
`
`In clinical trials lasting ~ 52 weeks, the use of sitagliptin
`and vildagliptin was well tolerated [48,19] . Withdrawal rates
`in patients randomized to these 2 DPP-4 inhibitors were
`similar to placebo. A recent meta-analysis suggested that
`the commonest adverse effects reported in slightly higher
`proportions of patients receiving sitagliptin or vildagliptin
`were nasopharyngitis (6.4 vs 6.1 o/o with placebo, risk ratio 1.2),
`urimuy tract infection (3.2 vs 2.4% with placebo, risk ratio 1.5)
`and headache (5.1 vs 3.9% with placebo, risk ratio 1.4) [23].
`Available data suggest that DPP-4 inhibitors are even better
`tolerated than metformin (39,48], glipizide [49], and acarbose [52].
`Hypoglycemia reported with the use of DPP-4 inhibitors
`
`was not severe, and was mainly evident when used in conjunc(cid:173)
`tion with SU. Thus, the proportions of patients reporting
`hypoglycemia were 12% (27 of 222) and 1.8% (4 of 219)
`in patients receiving sitagliptin plus glimepiride versus patients
`receiving glimep iride plus placebo, respectively (42]. In the
`postmarketing period, few serious cases of hypersensitivity
`reactions were reported possibly related to sitagliptin, including
`anaphylaxis, angiedema, and Stevens-Johnson syndrome [53].
`
`G. Exenatide versus DPP-4 inhibitors
`
`inhibitors sitagliptin and
`the 2 DPP-4
`Exenatide and
`vildagliptin
`share
`several
`features
`in common. Fi rst,
`reductions in HbA 1 c levels with the use of these agents are
`maximal at 12 -
`16 weeks, then reach a plateau. This
`observation is not consistent with the in-vitro findings,
`which suggest that exenatide [54] and sitagliptin
`[55 ) may
`promote regeneration or prevent apoptosis of pancreatic ~ cells,
`and therefore might virtually reverse the progressive course
`of diabetes. Nevertheless, long-term studies of several years
`are required to clarify this issue. Second, the reduction in
`HbAlc values
`is generally consistent irrespective of the
`class of background anti-diabetic therapy. Third, like other
`anti-diabetic agents, the magnitude of HbAl c reduction
`depends on the baseline levels, being greater with higher
`baseline HbAlc values. Fourth, as expected from
`their
`mechanism of action based on the incretin effect, these
`to a greater
`agents
`reduce postprandial hyperglycemia
`extent than fasting hyperglycemia [19,31]. Fifth, the risk of
`hypoglycemia is increased when these agents are used in
`conjunction with a SU [17,42). Adjustment of the dose of SU
`is therefore recommended when used concomitantly with
`these drugs. Meanwhile,
`there are important differences
`between exenatide and DPP-4 inhibitors exempl ifi ed by
`sitagliptin, as outlined in Table 1.
`
`7. Expert opinion
`
`There is no doubt that incretin-based drugs represent a useful
`add ition to the existing armamentarium of anti-diabetic
`drugs. These agents have several advantages. First, because of
`their distinct mechanism of action, they generally exert a
`beneficial effect on glycemic cont rol, irrespective of the type
`of background oral agents. Second, by targeting postprandial
`hyperglycemia more rhan fasting or pre-meal hyperglycemia,
`they complement the action of metfonnin, TZD, and long(cid:173)
`acting SU, which act mainly by lowering fasting plasma
`glucose. A tl1ird advantage is the mild, and possibly progressive,
`weight loss caused by exenatide, and the weight-neutral effect
`of the DPP-4 inhibitors. Fourth, the use of incretin-relared
`agents is uncommon ly associated with severe hypoglycemia.
`Moreover,
`the use of DPP-4 inhibitors is simple, with
`once-daily oral
`dosing
`irrespective of meal
`intake.
`Furrl1ermore, sitagliptin can be used in reduced dosage in rena l
`insufficiency (see Table 1).
`
`Th i~Wi~rRRir,Jf1'f,W~i ~ugs (2008) 17(6)
`atth.e· NLM and may be
`Su l:lj e ct U:S Co~y r ight Laws
`
`849
`
`Mylan EX 1003, Page 5
`
`

`
`Remarks
`
`Dose is changed in renal
`insufficiency (see below)
`
`Trials using exenatide in
`conjunction with insulin are
`underway
`
`No head-to-head t rials are
`avai lable comparing exenatide
`with sitagliptin
`
`CX)
`V1
`0
`
`Table 1. Comparison of exenatide with sitagliptin.
`
`Class of drug
`
`GLP-1 analog (i ncretin mimetic)
`
`Exenatide (Byetta)
`
`~
`~
`Ill -f~
`....,
`~ rt ::7'5"
`g_ g. Vi·:_
`m m 3 ::>
`~ :Z Ill~
`c r
`,........,.
`V) s: ~<!.5·
`S ru §:o
`~ ii ~s
`::::!. 3 ~VI
`((:} ~ s§
`rt Q"' 1;J 0
`:;;-Ill m·$
`~
`a. ~
`"'
`.... §
`
`Method of admi nistration
`
`Dose
`
`Indicat ions approved by the
`FDA for use in type 2 diabetes
`
`Subcutaneous injection w it hin
`the 60-min period before the
`morning and even ing meals
`
`5 iJg b.i.d. to be increased
`to 10 iJg b.i.d. after 4 weeks
`if t olerated
`
`Add-on t herapy to metformin,
`a combination of metformin
`and a SU, or a combination of
`metformin and a TZD
`
`Approxi mate average HbA 1 c
`reduction vs baseline
`
`-0.8%
`
`Effect on body weight
`
`Effect on satiety
`
`Main adverse effects
`
`W ithd rawal rates due
`to adverse effects
`
`Use in renal insufficiency
`
`Average weight loss
`of- 2 kg vs baseline
`
`May induce earl y satiety [59]
`
`Nausea (45- 51% vs 7 - 23%
`with placebo) and vomiting
`(12 - 14% vs 2 - 4%
`with placebo) [17-19]
`
`Higher than placebo
`(7 - 10% vs 1 - 5%), t han insulin
`glargine (1 0 vs < 1 %) [24]. and
`biphasic insulin aspart (8 vs 0%) [25]
`
`No dose adjustment is required in
`mild-to-moderate renal impairment
`(creat inine clearance 30 - 80 ml!min)
`Not recommended in more severe
`renal failure [60]
`
`Sitagliptin (Januvia)
`
`DPP-4 inhibitor
`(incretin enhancer)
`
`Oral tablet once daily,
`irrespective of meal t ime
`
`100 mg daily
`
`Monotherapy
`Adjunctive therapy to metformin,
`a TZD, a SU, or a combination
`of metformin and a SU, or
`metformin and TZD
`
`-0.7%
`
`In general, has neutral
`effect on weight
`
`No clear effect on satiety [1 1
`
`Nausea and vomit ing overall
`similar to placebo [31-34]
`
`Similar w it h sitag lipti n vs
`placebo [31 -34]. and possi bly
`lower with sitagliptin than
`metformin [39] and glipizide [49]
`
`In cases of moderate renal
`dysfunction (creatinine clea rance
`30- 49 ml/min or approximate serum
`creatinine levels 1.8 - 3 mg/dl in men,
`and 1.6 - 2.5 mg/dl in women), it is
`recommended to decrease the dose
`to 50 mg once daily. In more advanced
`renal dysfunction and patients on hemodialysis,*
`the recommended dose is 25 mg once daily [53]
`
`Approximate monthly cost in the USA
`
`$240
`
`$ 145
`
`*Sitagliptin is removed to a limited extent by hemodialysis. Therefore, the drug can be administered without reference to the timing of hemodialysis [61 ].
`
`tD
`
`::l n ..,
`tD =· ::l
`3
`3'
`tD ....
`;::;·
`"' Ql
`::l c.
`c.
`-a·
`"C ....
`0:
`'S.
`"C
`tD
`....
`"C
`0:
`Ql
`"' tD
`.;.
`::l :::r
`C"
`;::;:
`0 ..,
`"' :r
`Q:
`::l
`;::;·
`~ ....
`..,
`iii'
`
`iii -0 .., ... :::r
`
`tD ....
`;;;
`Ill ...
`3
`::l ...
`tD
`
`0 -~
`
`"C
`tD
`IV
`c.
`iii'
`C"
`tD ....
`"'
`
`tD
`
`Mylan EX 1003, Page 6
`
`

`
`Mikhail
`
`inhibitors
`Meanwhile, exenatide and current DPP-4
`have important limitations. First, it should be emphasized
`that > 50% of patients in clinical trials failed to achieve
`HbA1c levels < 7.0%, a fact that underscores their limited
`efficacy and the difficulty of optimizing diabetes control.
`Second, exenatide has to be injected twice daily, and is associated
`with high rates of nausea and vomiting. Third, while the
`short-term (~ 1 year) safety profile of two DPP-4 inhibitors -
`sitagliptin and vildaglitin - is reassuring, there are still so me
`unresolved issues related to their safety. For instance, the
`enzyme DPP-4 plays an important role in the immune system,
`being a T-cell co-stimulator [56]; this raises concern about
`possible immune suppression as result of DPP-4 inhibition.
`Second, in addition ro GLP-1 and GIP, DPP-4 inhibits the
`degradation of other peptides in vitro, such as substance P [56].
`Thus, there is a possibility that serum levels of such peptides
`may rise with the use of DPP-4 inhibitors leading to potential
`undesired effects. Third, there are two other enzymes, DPP-8
`and DPP-9, structurally related to DPP-4 but with largely
`[56]. Although in-vitro data suggest
`unknown functions
`that DPP-4 inhibitors display high selectivity for DPP-4, no
`in-vivo data are available. Finally, while the vast majority
`of trials of incretin-based agents were double-blinded and
`generally of good quality, all of them were sponsored by the
`corresponding manufacturer, and therefore may be open to
`different bias. The evaluation of this new class of anti-diabetic
`drugs will not be satisfactory until their efficacy and safety
`are car