Update
`
`Metformin: An Update
`
`Dmitri Kirpichnikov, MD; Samy I. McFarlane, MD; and James R. Sowers, MD
`
`Metformin is an insulin-sensitizing agent with potent antihyper-
`glycemic properties.
`Its efficacy in reducing hyperglycemia in
`type 2 diabetes mellitus is similar to that of sulfonylureas, thia-
`zolidinediones, and insulin. Metformin-based combination therapy
`is often superior to therapy with a single hypoglycemic agent. The
`antihyperglycemic properties of metformin are mainly attributed to
`suppressed hepatic glucose production, especially hepatic glu-
`coneogenesis, and increased peripheral tissue insulin sensitivity.
`
`Although the precise mechanism of hypoglycemic action of met-
`formin remains unclear, it probably interrupts mitochondrial oxi-
`dative processes in the liver and corrects abnormalities of intra-
`cellular calcium metabolism in insulin-sensitive tissues (liver,
`skeletal muscle, and adipocytes) and cardiovascular tissue.
`
`Ann Intern Med. 2002;137:25-33.
`For author affiliations, see end of text.
`
`www.annals.org
`
`Insulin resistance contributes greatly to development of
`
`cardiovascular disease in patients with the metabolic syn-
`drome and its extreme presentation, type 2 diabetes melli-
`tus. Therefore, treatment with an insulin-sensitizing agent,
`such as metformin, in patients with type 2 diabetes melli-
`tus may correct several of the primary pathophysiologic
`abnormalities of the metabolic syndrome. In diabetic pa-
`tients, metformin appears to provide cardiovascular protec-
`tion that cannot be attributed only to its antihyperglycemic
`effects. These additional cardioprotective effects in these
`patients may be related to the favorable actions of met-
`formin on lipid metabolism, vascular smooth-muscle and
`cardiomyocyte intracellular calcium handling, endothelial
`function, hypercoagulation, and platelet hyperactivity. We
`discuss known mechanisms by which metformin exerts its
`beneficial glycemic and cardiovascular actions.
`
`CLINICAL ROLE OF METFORMIN
`Metformin, an insulin-sensitizing biguanide used to
`treat type 2 diabetes, has been shown to be as effective as
`insulin or sulfonylureas when used as monotherapy (1–5).
`In conjunction with diet, metformin reduces fasting glu-
`cose concentration by 2.78 to 3.90 mmol/L (50 to 70
`mg/dL), which corresponds to a 1.3% to 2.0% reduction
`in hemoglobin A1c values (1, 2, 4, 6 – 8). The magnitude of
`plasma glucose reduction is related to pretreatment glucose
`levels (7, 9). The efficacy of metformin monotherapy has
`been shown to be independent of age, body weight, eth-
`nicity, duration of diabetes, and insulin and C-peptide lev-
`els (1, 2).
`Metformin may have special benefits in overweight
`patients with type 2 diabetes. Unlike sulfonylureas, insulin,
`and thiazolidinediones, metformin does not affect body
`mass index (1) or decreases body weight in obese patients
`with (4, 10) and without (11, 12) diabetes. Significant
`reductions in total body fat and visceral fat have been ob-
`served in women with preexistent abdominal or visceral
`obesity who are treated with metformin (11). Excessive fat
`localized to the paraintestinal region is a major contributor
`to the pathogenesis of the cardiovascular metabolic syn-
`drome (13, 14), and the reduction in visceral fat (second-
`
`ary to weight loss or fat redistribution) may have additional
`cardiovascular benefits in insulin-resistant persons treated
`with metformin (13, 14). Weight loss during metformin
`treatment has been attributed to decreased net caloric in-
`take (15), probably through appetite suppression, an effect
`that is largely independent of gastrointestinal side effects of
`metformin (such as nausea and diarrhea) (10). Reduction
`in hyperinsulinemia related to reduced insulin resistance
`may have an additive effect on weight reduction in obese
`insulin-resistant persons (13, 14).
`At doses of 500 to 1500 mg, metformin has an abso-
`lute oral bioavailability of 50% to 60% (16). The drug is
`not protein bound and therefore has a wide volume of
`distribution (8), with maximal accumulation in the small-
`intestine wall (17). Metformin undergoes no modifications
`in the body and is secreted unchanged by rapid kidney
`excretion (through glomerular filtration and, possibly, tu-
`bular secretion) (8). Impaired kidney function slows elim-
`ination and may cause metformin accumulation (18). The
`H2-blocker cimetidine competitively inhibits renal tubular
`secretion of metformin, significantly decreasing its clear-
`ance and increasing its bioavailability (16, 19).
`
`METFORMIN AS A PART OF COMBINATION THERAPY
`Metformin has been shown to be effective in combi-
`nation with insulin, sulfonylureas (2, 10, 20, 21), and thia-
`zolidinediones (22). This finding is important because
`single-drug therapy often fails to maintain normoglycemia,
`particularly as diabetes progresses (23, 24). As seen in the
`United Kingdom Prospective Diabetes Study (UKPDS),
`50% of patients treated with diet or a single antidiabetic
`drug achieved the target hemoglobin A1c value of less than
`7% after 3 years of follow-up; after 9 years, only 25%
`maintained this goal (24). As diabetes progresses and treat-
`ment with maximal doses of sulfonylurea fails, addition of
`metformin significantly improves glycemic control (2). In
`the UKPDS trial, combination therapy tended to control
`glycemia more effectively than monotherapy (hemoglobin
`A1c value, 0.075 [7.5%] versus 0.081 [8.1%]) (23).
`© 2002 American College of Physicians–American Society of Internal Medicine E-25
`
`Downloaded From: http://annals.org/ by a Reprints Desk User on 07/06/2015
`
`MYLAN Ex. 1015, Page 1
`
`

`
`Update Metformin: An Update
`
`PRACTICAL CONSIDERATIONS IN METFORMIN THERAPY
`The ideal patient for initiation of metformin treatment
`would be an obese person with type 2 diabetes mellitus who
`has normal kidney function (creatinine concentration ⬍133
`␮m d/L [⬍1.5 mg/dL] in men and ⬍124 ␮m d/L in women,
`or creatinine clearance ⬎1.17 mL/s without coexistent symp-
`tomatic congestive heart failure or a hypoxic respiratory con-
`dition) (9, 16, 25). Contraindications to metformin therapy
`are liver failure, alcoholism, and active moderate to severe in-
`fection (9, 25); these conditions predispose to development of
`lactic acidosis, either by increased production or decreased
`metabolism of lactic acid (9, 16–18, 25). Administration of
`radiocontrast material to a patient with diabetes may worsen
`already-compromised kidney function and cause accumula-
`tion of metformin, leading to toxic levels of drug. Further-
`more, administration of general anesthesia may cause hypo-
`tension, which leads to renal hypoperfusion and peripheral
`tissue hypoxia with subsequent lactate accumulation (25–28).
`Therefore, if administration of radiocontrast material is re-
`quired or urgent surgery is needed, metformin should be
`withheld and hydration maintained until preserved kidney
`function is documented at 24 and 48 hours after the interven-
`tion (9, 26–28). Metformin should be used with caution in
`elderly patients, whose reduced lean body mass may lead to
`misleading low creatinine concentrations that fail to reflect
`decreased glomerular filtration rates (9, 25–28).
`Metformin therapy should be initiated with a single
`dose of medication (usually 500 mg) taken with the pa-
`tient’s largest meal to prevent gastrointestinal symptoms.
`Gastrointestinal symptoms generally disappear within 2
`weeks of treatment (10, 11). Medication doses may be
`increased by 500-mg increments every 1 to 2 weeks, as
`indicated by glycemic control, until a desirable blood glu-
`cose level or the maximal recommended daily metformin
`dose of 2550 mg is reached (2, 25). The hypoglycemic
`effect of metformin is dose related, and a plateau of hypo-
`glycemic action is achieved at a daily dose of 2000 mg (6).
`Side effects of metformin are mostly limited to diges-
`tive tract symptoms, such as diarrhea, flatulence, and ab-
`dominal discomfort (1, 6, 8 –10). These symptoms are
`dose dependent and can usually be avoided by slow titra-
`tion and, in some cases, reduction of the dose (9). About
`5% of patients cannot tolerate treatment because of gastro-
`intestinal side effects (6, 9, 10). The mechanisms of these
`gastrointestinal side effects remain unclear but probably are
`related to accumulation of high amounts of metformin in
`the intestinal tissue (17), with subsequent elevation of local
`lactate production. Histologic examination has not re-
`vealed changes in the intestinal mucosa in metformin-
`treated animals (29), indicating a functional rather than a
`structural basis for gastrointestinal symptoms. Ten percent
`to 30% of patients receiving long-term metformin therapy
`develop vitamin B12 malabsorption, as indicated by de-
`creased concentrations of total vitamin B12 and its bioavail-
`able form, holotranscobalamin (2, 30). Metformin inter-
`
`E-26 2 July 2002 Annals of Internal Medicine Volume 137 • Number 1
`
`feres with mucosal-cell intracellular calcium handling, thus
`disrupting calcium-dependent absorption of vitamin B12 in
`the ileum (30). Such decreases in vitamin B12 levels rarely
`have clinical significance (2, 9).
`Development of hypoglycemia during metformin
`monotherapy is rare because metformin only partially sup-
`presses gluconeogenesis in the liver and does not stimulate
`insulin production (9, 31).
`Lactic acidosis is a life-threatening complication of bigua-
`nide therapy that carries a mortality rate of 30% to 50% (28).
`Metformin therapy may increase blood lactate levels (1) and is
`occasionally associated with development of lactic acidosis (2,
`28). The estimated incidence of metformin-associated lactic
`acidosis is 0.03 cases per 1000 patient-years (25), which is 10
`to 20 times lower than that seen with phenformin therapy
`(28). Development of lactic acidosis appears to be unrelated to
`plasma metformin concentrations (28), and even in persons
`with chronic renal
`insufficiency, metformin accumulation
`does not necessarily lead to lactic acidosis (18). Development
`of lactic acidosis is almost always related to coexistent hypoxic
`conditions that are probably responsible for the associated
`high mortality rate. In one report, 91% of patients who de-
`veloped lactic acidosis while being treated with metformin had
`a predisposing condition, such as congestive heart failure, re-
`nal insufficiency, chronic lung disease with hypoxia, or age
`older than 80 years (26). Thus, patients with compromised
`renal function or coexistent hypoxic conditions should not be
`given metformin. Chronic or acute intake of large amounts of
`alcohol may potentiate the effect of metformin on lactate me-
`tabolism. A careful history of alcohol use is therefore impor-
`tant before starting metformin therapy (26, 27).
`
`MECHANISMS OF ANTIHYPERGLYCEMIC ACTION OF
`METFORMIN
`The glucose-lowering effects of metformin are mainly a
`consequence of reduced hepatic glucose output (primarily
`through inhibition of gluconeogenesis and, to a lesser extent,
`glycogenolysis) and increased insulin-stimulated glucose up-
`take in skeletal muscle and adipocytes (25, 27, 31–35) (Figure
`1). Its major mode of action is to reduce hepatic glucose pro-
`duction, which is increased at least twofold in patients with
`type 2 diabetes (32, 36). In a recent study of the mechanism
`by which metformin decreases endogenous glucose produc-
`tion in patients with type 2 diabetes, the increased plasma
`glucose level was attributed to a threefold increase in the rate
`of gluconeogenesis, as assessed by nuclear magnetic resonance
`spectroscopy (32). Metformin treatment decreased fasting
`plasma glucose concentrations by 25% to 30% and reduced
`glucose production (32), findings that are consistent with
`those of other investigators (27, 35). The decrease in glucose
`production was attributable to a reduction in the rate of glu-
`coneogenesis (32).
`Data from in vivo studies (27, 32, 36) are consistent with
`those of in vitro studies demonstrating an inhibitory effect of
`metformin on gluconeogenesis (37, 38) (Figure 1). For exam-
`
`www.annals.org
`
`Downloaded From: http://annals.org/ by a Reprints Desk User on 07/06/2015
`
`MYLAN Ex. 1015, Page 2
`
`

`
`ple, metformin was observed to decrease gluconeogenesis in
`perfused liver, primarily through inhibition of hepatic lactate
`uptake (37). Others reported that metformin therapy de-
`creased concentrations of adenosine triphosphate in isolated
`rat hepatocytes (38). Because adenosine triphosphate is an
`allosteric inhibitor of pyruvate kinase, the investigators sug-
`gested that the metformin-mediated reduction in hepatic glu-
`cose production resulted from increased pyruvate kinase flux.
`Metformin also decreases gluconeogenic flux through inhibi-
`tion of pyruvate carboxylase–phosphoenolpyruvate carboxyki-
`nase activity and possibly through increased conversion of
`pyruvate to alanine (34). Metformin also facilitates insulin-
`induced suppression of gluconeogenesis from several sub-
`stances, including lactate, pyruvate, glycerol, and amino acids
`(31), and opposes the gluconeogenic actions of glucagon (39)
`(Figure 1).
`The exact mechanism through which metformin re-
`duces hepatic glucose production remains unclear, but its
`primary site of action appears to be hepatocyte mitochon-
`dria, where it disrupts respiratory chain oxidation of com-
`plex I substrates (for example, glutamate) (15, 39). Inhibi-
`tion of cellular respiration decreases gluconeogenesis (39)
`and may induce expression of glucose transporters and,
`therefore, glucose utilization (40). It is not clear whether
`metformin acts on mitochondrial respiration directly by
`slow permeation across the inner mitochondrial membrane
`(39) or by unidentified cell-signaling pathways (15). It has
`been suggested that biguanides bind specifically and com-
`petitively to divalent cation sites on proteins, thus interfer-
`ing with intracellular handling of calcium ([Ca2⫹]i) (41,
`42) especially in the mitochondria (41). Davidoff and col-
`leagues (41) showed that even small doses of biguanides
`increase the rates of [Ca2⫹]i uptake in isolated hepatic
`mitochondria, where [Ca2⫹]i serves as a potent activator of
`respiration (Figure 1). This effect was
`mitochondrial
`shown at biguanide concentrations as low as 5 to 10 ␮m
`(41), levels that are expected in the liver with antihypergly-
`cemic doses of the drug and are 20- to 50-fold lower than
`those that inhibit mitochondrial respiration. In several tis-
`sues, including skeletal muscle and adipocytes, metformin
`facilitates trafficking of glucose transporters 4 and 1 to the
`plasma membrane (25, 31, 43). Moreover, metformin may
`increase the glucose transport capacity of glucose trans-
`porter 4, and to some extent, glucose transporters 1 (31).
`The effects of metformin on peripheral insulin-sensi-
`tive tissues require the presence of insulin for its full action.
`Metformin enhances most of the biological actions of in-
`sulin, including glucose transport and glycogen and lipid
`synthesis,
`in persons with preexisting insulin resistance
`(31). It facilitates glucose transport in cultured skeletal
`muscle in the absence of insulin (44, 45). Metformin acti-
`vates insulin and tyrosine kinase activity in insulin-like
`growth factor-1 receptor of vascular smooth-muscle cells
`independently of insulin action (46). The drug activates
`tyrosine kinase in Xenopus oocytes, with subsequent stim-
`ulation of inositol 1,4,5-triphosphate production and gly-
`
`www.annals.org
`
`Metformin: An Update
`
`Update
`
`Figure 1. Mechanisms of metformin action on hepatic glucose
`production and muscle glucose consumption.
`
`Metformin decreases hepatic gluconeogenesis by interfering with respi-
`ratory oxidation in mitochondria. It suppresses gluconeogenesis from
`several substrates, including lactate, pyruvate, glycerol, and amino acids.
`In addition, metformin increases intramitochondrial levels of calcium
`(Ca⫹⫹), a modulator of mitochondrial respiration. In insulin-sensitive
`tissues (such as skeletal muscle), metformin facilitates glucose transport
`by increasing tyrosine kinase activity in insulin receptors and enhancing
`glucose transporter (GLUT) trafficking to the cell membrane. ADP ⫽
`䢇䢇䢇; ATP ⫽ 䢇䢇䢇; Ca⫹⫹ ⫽ intracellular calcium levels; OAA ⫽
`䢇䢇䢇; PEP ⫽ phosphoenolpyruvate; Pi ⫽ 䢇䢇䢇; TK ⫽ 䢇䢇䢇.
`
`cogen synthesis (47) (Figure 1). Thus, metformin has met-
`abolic
`effects on insulin-sensitive
`tissues
`that may
`contribute to its glucose-lowering effect.
`Metformin has been shown to reduce free fatty acid
`oxidation by 10% to 30% (25, 31–33). Elevated levels of
`free fatty acid are commonly seen in diabetes and obesity
`(48), and they contribute to increased hepatic glucose pro-
`duction and development of insulin resistance (49, 54)
`(Figure 2). Increased fatty acid oxidation inhibits key en-
`zymes of the glycolytic pathway by accumulation of acetyl
`coenzyme A and citrate, by-products of free fatty acid ox-
`idation (51). Increased glucose 6-phosphate concentra-
`tions, in turn, inhibit the hexokinase enzyme, resulting in
`reduced glucose uptake and oxidation (51). In addition,
`free fatty acid independently inhibits insulin receptor sub-
`strate-1–associated PI3-kinase activity (52) and subse-
`quently attenuates transmembrane glucose transport (48)
`(Figure 2). By decreasing free fatty acid levels, metformin
`not only improves insulin sensitivity but may also help
`correct impaired insulin secretion by ␤-cells (53). Met-
`formin has no direct effect on ␤-cell function (9), but it
`can improve insulin secretion that has been altered by
`long-term exposure to free fatty acid or hyperglycemia
`(glucose toxicity) (53).
`Metformin may also improve hyperglycemia by attain-
`ing high concentrations in the small intestine (17, 31) and
`decreasing intestinal absorption of glucose (29, 54), an ac-
`tion that may contribute to decreased postprandial blood
`
`2 July 2002 Annals of Internal Medicine Volume 137 • Number 1 E-27
`
`Downloaded From: http://annals.org/ by a Reprints Desk User on 07/06/2015
`
`MYLAN Ex. 1015, Page 3
`
`

`
`Update Metformin: An Update
`
`Figure 2. Metformin and fatty acids.
`
`Metformin inhibits fatty acid (FA) production and oxidation, thereby reducing fatty acid–induced insulin resistance and hepatic glucose production.
`CoA ⫽ coenzyme A; CPT ⫽ 䢇䢇䢇; FFA ⫽ free fatty acid; GLUT ⫽ glucose transporter; IGF-1 ⫽ 䢇䢇䢇; IRS-1 ⫽ 䢇䢇䢇; OAA ⫽ 䢇䢇䢇; PDH ⫽
`䢇䢇䢇; PFK ⫽ 䢇䢇䢇; PI-3 ⫽ 䢇䢇䢇.
`
`glucose levels (55). It has been speculated that increased
`glucose consumption in the small intestine of metformin-
`treated patients may prevent further glucose transport to
`the hepatic circulation (29).
`In summary, metformin decreases hepatic glucose pro-
`duction through inhibition of gluconeogenesis and possi-
`bly glycogenolysis and improves peripheral insulin sensitiv-
`ity.
`In addition, metformin decreases gastrointestinal
`glucose absorption and indirectly improves pancreatic
`␤-cell response to glucose by reducing glucose toxicity and
`free fatty acid levels.
`
`EFFECT OF METFORMIN IN THE POLYCYSTIC OVARY
`SYNDROME
`Hyperinsulinemia reflecting insulin resistance is a com-
`mon feature in lean and obese patients with the polycystic
`ovary syndrome (11, 56, 57). Hyperinsulinemia contributes
`directly to excessive testosterone production by the ovaries
`(56) and decreased synthesis of sex hormone–binding globu-
`lin in the liver (11, 58), thereby increasing levels of total and
`free testosterone. Metformin therapy increases insulin sensitiv-
`ity and decreases insulin levels in patients with the polycystic
`ovary syndrome (56, 57, 59). Improvement of hyperinsulin-
`emia is associated with decreased levels of total and free tes-
`tosterone (11, 12, 57, 59) and increased estradiol (12) levels.
`
`E-28 2 July 2002 Annals of Internal Medicine Volume 137 • Number 1
`
`Clinically, administration of metformin improves hirsutism
`(11), normalizes menstrual cycles (11, 12, 57, 59), and in-
`duces ovulation (57, 59) in a substantial number of patients
`with the polycystic ovary syndrome.
`
`EFFECT OF METFORMIN TREATMENT ON
`CARDIOVASCULAR MORBIDITY AND MORTALITY
`In the UKPDS 34, metformin therapy was compared
`with conventional treatment or treatment with sulfonyl-
`urea or insulin (5). In this trial, which was designed to
`achieve fasting plasma glucose levels less than 6 mmol/L
`(⬍108 mg/dL), 342 patients with newly diagnosed type 2
`diabetes were allocated to receive metformin treatment and
`951 patients were allocated to receive either chlorpropam-
`ide, glibenclamide, or insulin. The control group included
`411 overweight diabetic patients who were randomly as-
`signed to conventional therapy, primarily with diet alone,
`which resulted in suboptimal glycemic control. During 10
`years of
`follow-up, both drug-treated groups achieved
`equal degrees of glycemic control (median hemoglobin A1c
`value of 0.074 [7.4%]), whereas the conventionally treated
`group had a median hemoglobin A1c value of 0.08 (8.0%)
`(5). Compared with the conventionally treated group,
`metformin-treated patients had a risk reduction of 32%
`any diabetes-related end point, 39% for myocardial infarc-
`
`www.annals.org
`
`Downloaded From: http://annals.org/ by a Reprints Desk User on 07/06/2015
`
`MYLAN Ex. 1015, Page 4
`
`

`
`tion, 42% for diabetes-related death, and 36% for all-cause
`mortality (5). These differences may be partially explained
`by differences in the degree of glycemic control between
`the metformin and diet groups. In the UKPDS 35 (60),
`the risk for cardiovascular events, stroke, and all-cause
`death was closely related to the degree of glycemia in dia-
`betic patients. In that study, each 1% reduction in the
`hemoglobin A1c value during treatment of type 2 diabetes
`was associated with a reduction of 21% in diabetes-related
`deaths, 14% in the incidence of myocardial infarction,
`12% in fatal and nonfatal strokes, and 16% in heart failure
`(60). Nevertheless, metformin was more effective than sul-
`fonylureas or insulin in reducing rates of any diabetes-
`related end point, all-cause mortality, and stroke, even
`though both agents decreased hemoglobin A1c values
`equally (5). These observations suggest that metformin
`might have additional cardiovascular protective actions be-
`yond its antihyperglycemic properties. However, data indi-
`cate that metformin in combination with sulfonylurea
`might increase cardiovascular mortality in patients with
`type 2 diabetes (5, 61). In those studies, metformin was
`not used as an initial therapy but rather was added to
`treatment when sulfonylurea therapy failed. Patients taking
`combination therapy with metformin and sulfonylurea
`tended to have long-standing poorly controlled diabetes
`before addition of the biguanide (62). Moreover, they had
`greater obesity (61), which could independently increase
`mortality. Therefore, the reported increase in risk for car-
`diovascular disease in patients treated with combination
`therapy might reflect selection bias attributable to the natural
`history of long-standing diabetes rather than to adverse effects
`of this combination.
`
`MECHANISM OF THE CARDIOPROTECTIVE ACTION OF
`METFORMIN
`Insulin resistance, a cornerstone of type 2 diabetes and
`the metabolic cardiovascular syndrome, is commonly asso-
`ciated with hypertension, abdominal obesity, atherogenic
`dyslipidemia, and vascular dysfunction, all of which con-
`tribute greatly to the development of accelerated athero-
`sclerosis (63). Hyperinsulinemia reflects insulin resistance
`and may be an independent risk factor for coronary artery
`disease (64 – 66). Metformin, an insulin-sensitizing agent,
`decreases insulin resistance in patients with (20, 31, 55)
`and without (11, 12, 57, 67) diabetes, thus effectively re-
`ducing baseline and glucose-stimulated insulin levels (12,
`20, 55, 57, 67).
`Several studies have shown that metformin improves
`lipoprotein profiles in diabetic patients (2, 10, 20, 55, 68).
`Dyslipidemia in diabetes is characterized by hypertriglycer-
`idemia (increased levels of very low-density lipoprotein
`cholesterol); decreased levels of high-density lipoprotein
`cholesterol; and elevated levels of small, dense atherogenic
`low-density lipoprotein cholesterol (LDL) particles. The
`increased levels of free fatty acid that occur in obesity and
`
`www.annals.org
`
`Metformin: An Update
`
`Update
`
`poorly controlled diabetes (48) contribute not only to de-
`velopment of insulin insensitivity but also to increased syn-
`thesis and secretion of very low-density lipoprotein (69).
`Elevated triglyceride levels inhibit degradation of apopro-
`tein B in the liver and lead to increased assembly of very
`low-density lipoprotein and smaller, denser LDL particles
`(69). Excessive generation of reactive oxygen species and
`free radicals (such as peroxynitrates) by cardiovascular tis-
`sue, in combination with increased nonenzymatic glycation
`of
`lipoproteins (glycooxidation),
`leads to formation of
`atypical glycooxidized LDL particles. These particles bind
`poorly to classic LDL receptors but have high affinity for
`“scavenger” receptors, which are located predominantly on
`macrophages (63). Accumulation of glycooxidized small,
`dense LDL particles converts macrophages into foam cells,
`which are essential participants in the early steps of athero-
`sclerotic plaque formation (63). Compared with the gen-
`eral population, diabetic persons have a twofold to fourfold
`increased risk for cardiovascular disease at any cholesterol
`level (70), which indicates a more aggressive type of dys-
`lipidemia. Furthermore, decreasing cholesterol and triglyc-
`eride levels has been shown to be particularly beneficial in
`patients with diabetes (70, 71). In addition, hypertriglyc-
`eridemia may be an independent risk factor for cardiovas-
`cular disease in patients with type 2 diabetes (72). Met-
`formin has major effects on lipid metabolism in patients
`with insulin resistance. It decreases plasma levels of free
`fatty acid (20, 73) and oxidation of these acids by tissue
`(25, 28, 32); it decreases levels of triglycerides (2, 10, 20,
`55, 74) and, therefore, very low-density lipoprotein (20).
`Metformin therapy decreases levels of total cholesterol (2,
`68, 74) and LDL cholesterol (2, 68, 74) while maintaining
`(68, 74) or increasing (2, 20, 55, 57, 67) levels of high-
`density lipoprotein cholesterol. Metformin decreases oxida-
`tive stress and reduces lipid oxidation (75) by lowering
`plasma glucose levels (2). Taken together, these observa-
`tions suggest that the beneficial effects of metformin on
`lipoprotein metabolism may contribute to its protective
`effects against cardiovascular disease.
`Metformin has also been shown to lessen hypercoagula-
`tion and increase fibrinolysis in insulin-resistant states by de-
`creasing levels of plasminogen activator inhibitor-1 (76, 77)
`and increasing tissue plasminogen activator activity (74).
`Therapy with metformin also reduces thrombogenic propen-
`sity by decreasing levels of tissue plasminogen activator anti-
`gen (78) and von Willebrand factor (78). In the Biguanides
`and the Prevention of the Risk of Obesity study, 457 nondi-
`abetic patients with visceral obesity (body mass index of 32.5
`kg/m2) were randomly assigned to treatment with diet or met-
`formin (850 mg twice daily) (78). Weight loss was associated
`with a 30% to 40% decrease in plasminogen activator inhib-
`itor-1 activity, regardless of the method used, whereas met-
`formin produced significantly larger decreases in von Wille-
`brand factor levels than did diet therapy (78). Furthermore,
`metformin therapy decreased platelet aggregation in diabetic
`patients treated with 1700 mg/d (79). Thus, metformin ther-
`
`2 July 2002 Annals of Internal Medicine Volume 137 • Number 1 E-29
`
`Downloaded From: http://annals.org/ by a Reprints Desk User on 07/06/2015
`
`MYLAN Ex. 1015, Page 5
`
`

`
`Update Metformin: An Update
`
`Table. Direct and Indirect Cardiovascular Protective Effects of
`Metformin Therapy
`
`Decreases hyperglycemia
`Improves diastolic function
`Decreases total cholesterol levels
`Decreases very low-density lipoprotein cholesterol levels
`Decreases low-density lipoprotein cholesterol levels
`Increases high-density lipoprotein cholesterol levels
`Decreases oxidative stress
`Improves vascular relaxation
`Decreases plasminogen activator inhibitor-1 levels
`Increases tissue plasminogen activator activity
`Decreases von Willebrand factor levels
`Decreases platelet aggregation and adhesion
`
`apy appears to lessen the hypercoagulability and exaggerated
`platelet aggregation and adhesion in diabetic patients (Table).
`
`METFORMIN AND DIABETIC CARDIOMYOPATHY
`Persons with diabetes have a high prevalence of con-
`gestive heart failure (80) secondary to diabetic, hyperten-
`sive, and ischemic changes in the myocardium. Diabetic
`cardiomyopathy, a unique clinical entity, is characterized
`by structural changes in the myocardium (fibrosis) and
`functional alterations in diastolic relaxation and ventricular
`compliance (81– 84) (Figure 3). Delayed diastolic relax-
`ation in diabetic cardiomyopathy is related to diminished
`removal of [Ca2⫹]i from cardiomyocytes after systolic con-
`traction (82, 83). Hyperglycemia has been shown to con-
`tribute to these functional changes (82– 84), and insulin
`resistance also directly contributes to these abnormalities
`(85). Metformin treatment of streptozotocin diabetic rats
`corrects these functional cardiac abnormalities (84, 86),
`perhaps through tyrosine kinase– dependent increases in
`intracellular [Ca2⫹]i removal after systole (84). This car-
`dioprotective action of metformin was shown to be insulin
`independent (84). Moreover, treatment of spontaneously
`hypertensive rats with metformin has been reported to de-
`crease heart rate (a sympathoinhibitory effect) more than
`placebo (87, 88). Although these findings are of interest,
`no clinical trials to date have investigated the effect of met-
`formin on the development and course of congestive heart
`failure in diabetic patients.
`
`METFORMIN AND VASCULAR REACTIVITY
`Hypertension is often associated with insulin resis-
`tance (89). Diabetic patients have a higher incidence of
`hypertension compared with the general population, and
`hypertensive persons are more prone to develop diabetes
`(90, 91). Recently, investigators demonstrated that defec-
`tive insulin signaling may contribute to increased vascular
`resistance (92), which is the hallmark of hypertension in
`type 2 diabetes (89). Insulin normally acts through the
`PI3-kinase pathway to activate nitric oxide synthase, en-
`hance sodium pump activity in vascular smooth muscle,
`and increase glucose transmembrane transport (63). More-
`
`E-30 2 July 2002 Annals of Internal Medicine Volume 137 • Number 1
`
`over, insulin is responsible for the normal handling of di-
`valent cations in vascular smooth muscle (92). Those pro-
`cesses are altered in insulin resistance. Impaired vascular
`insulin action may result
`in impaired nitric oxide–
`dependent vascular relaxation, decreased sodium pump ac-
`tivity, and increased levels of [Ca2⫹]i in vascular smooth
`muscle in patients with type 2 diabetes (14, 63, 92). These
`abnormalities in divalent cation and nitric oxide metabo-
`lism appear to play a role in the increased vascular resis-
`tance and impaired vasorelaxation that characterize hyper-
`tension, which frequently occurs in diabetic patients (92).
`Several reports indicate an antihypertensive effect of
`metformin in animals (88, 93–95) and humans (74, 96).
`In contrast, no effect of metformin on blood pressure was
`reported in other human studies (1, 2, 23). Careful 24-
`hour ambulatory studies may better characterize the effects
`of metformin on blood pressure in diabetic patients (89).
`Potential mechanisms of antihypertensive action of met-
`formin are complex and include both insulin-dependent
`and insulin-independent vasodilatory actions (Figure 3).
`Acute administration of metformin to rat tail arteries in-
`creases repolarization and causes subsequent artery relax-
`ation (97) through reduction in agonist-induced increase
`in intracellular levels of [Ca2⫹]i vascular smooth muscle
`(46, 94). This attenuation of [Ca2⫹]i responses may be
`secondary to increased nitric oxide production by vascular
`smooth muscle during exposure to metformin (94). In-
`deed, nitric oxide has been shown to decrease vascular
`smooth muscle [Ca2⫹]i responses to vasoconstrictor ago-
`nists through activation of the cyclic guanosine monophos-
`phate pathway (98). Metformin may also reduce [Ca2⫹]i
`
`Figure 3. Proposed cellular mechanisms of metformin action in
`the vascular smooth-muscle cells and cardiomyocytes.
`
`In vascular smooth-muscle cells, metformin decreases vasoconstriction by
`enhancing sodium pump activity and nitric oxide (NO) production,
`causing a decrease in intracellular calcium levels (Ca⫹⫹). Metformin
`improves diastolic relaxation by enhancing calcium removal from cardio-
`myocytes after systole. ATP ⫽ 䢇䢇䢇; CGMP ⫽ 䢇䢇䢇; GTP ⫽
`䢇䢇䢇; K-ATP ⫽ 䢇䢇䢇.
`
`www.annals.org
`
`Downloaded From: http://annals.org/ by a Reprints Desk User on 07/06/2015
`
`MYLAN Ex. 1015, Page 6
`
`

`
`by incre