What are the antihyperglycaemic mechanisms of metformin

Type 2 diabetes mellitus

The antihyperglycemic therapy of type 2 diabetes mellitus is complex and always presents a challenge. According to current German guidelines, the therapy goal is usually an HbA1c value below 6.5% while avoiding hypoglycaemia. For this purpose, based on a lifestyle change, oral antidiabetic therapy and insulin therapy are used alone or in combination. This article gives an overview of the mode of action, the effectiveness and the side effects of the available drug therapy forms on the basis of the available study data. New pharmacological therapeutic approaches are also briefly presented.
Drug Therapy 2007; 25: 175-86.

Glycemic Control Target Values

Type 2 diabetes mellitus is a very common disease with deleterious consequences in Europe, with a prevalence of at least 8 to 10% in those over 40 years of age [1]. There is consensus that the aim of anti-diabetic therapy in patients with normal life expectancy must be blood sugar adjustment (glycemic control) close to the norm; However, this glycemic control is defined differently in the multitude of local, national and international guidelines that have now existed. Before considering the individual forms of therapy, the definition of the therapy goal should therefore first be discussed in more detail. Regardless of this, it must be pointed out that, on the basis of endpoint studies that have so far only been sporadic, a better cardiovascular prognosis can be assumed with anti-hyperglycaemic settings close to the norm [2].

In the current national recommendations of the German Diabetes Society (DDG) of May 2006 (DDG Practice Guideline, [3]), adequate glycemic control in type 2 diabetes mellitus is based on preprandial or fasting blood glucose values ​​of 80 to 120 mg / dl and, on the other hand, an HbA1c value of less than 6.5% [4]. Although not explicitly formulated in the practice guideline, blood sugar values ​​are usually determined by self-measurement (SMBG = self monitoring of blood glucose) from capillary whole blood. It must be taken into account that fasting and preprandial blood sugar values ​​in capillary whole blood are around 15 mg / dl lower than the corresponding values ​​from venous plasma, to which the common diagnostic limit value for a diabetic metabolic state of 126 mg / dl refers [ 5]. The current version of the DDG practice guideline [3] does not make any direct reference to the postprandial, that is, blood sugar values ​​measured two hours after eating. At best, these are taken into account indirectly, since it can be assumed that an HbA1c value of below 6.5% cannot be achieved without a strict adjustment of the postprandial blood sugar level as well [6–8].

In addition to the DDG practice guideline [3] as the most current recommendation at national level, there is also the care guideline for type 2 diabetes mellitus from 2002 (NVL T2DM 2002, [9, 10]). It is revised module by module, the first two modules “Retinal Complications” and “Foot Complications” were published in October and November 2006 [136]. The 2002 Care Guideline is a consensus that was developed under the patronage of the German Medical Association (BÄK) and the support of the Working Group of Scientific Medical Societies (AWMF) and the National Association of Statutory Health Insurance Physicians (KBV) of specialist societies active in diabetes (Drugs Commission of the German Medical Association = AkdÄ, German Society for Internal Medicine = DGIM, German Diabetes Society = DDG, Special Commission for Diabetes Saxony = FDS), has been developed [9, 10]. In the comprehensive recommendations for the treatment of type 2 diabetes mellitus set out therein, on the one hand a preprandial or fasting glucose value of 80 to 100 mg / dl in capillary whole blood, on the other hand an HbA1c value below 6.5% and a postprandial blood sugar value between 80 and 135 mg / dl (for the postprandial blood sugar level, in contrast to the fasting blood sugar level, there are no relevant differences between capillary whole blood and venous plasma) were defined as targets. The recommendation for the postprandial blood sugar level with its lower interval limit of 80 mg / dl appears to be quite strict (and therefore not very realistic), so that the revised version of the national supply guideline should be awaited here.

For specific recommendations on postprandial blood sugar values, two internationally recognized guidelines, namely the American College of Endocrinology (ACE) from 2002 [11] and the International Diabetes Federation (IDF) from 2005 [12], referenced, in which 140 mg / dl and 145 mg / dl are defined as target values. Otherwise, the guidelines for the fasting blood sugar value measured by capillary means (<110 mg / dl) and the HbA1c value (≤ 6.5%) are essentially the same as the DDG practice guideline.

Finally, the American Diabetes Association (ADA) as the world's most influential diabetes specialist society, in its current recommendations from January 2006, defined the target values ​​for the HbA1c value with generally below 7.0% and for the postprandial blood sugar level of below 180 mg / dl the aforementioned professional societies [13]. When it comes to the fasting or preprandial blood sugar value (interval between 90 and 130 mg / dl), the ADA is more liberal than the aforementioned bodies.

General therapy scheme for type 2 diabetes mellitus

After the initial diagnosis of type 2 diabetes mellitus - based on the pathophysiology of the disease - a fundamental change in lifestyle (summarized in detail [14–16]) is the first therapeutic goal. If after three months, despite attempts to reduce weight and increase physical activity, the HbA1c value is still above 7%, the current guidelines of the DDG (modified shown in Fig. 1) see the addition of oral antidiabetic drugs initially in mono-, then in combination therapy and finally insulin therapy [4]. These therapeutic steps are presented below.

Oral anti-diabetic therapy

The preparations currently approved for this indication generally work through three basic principles, namely through

  • Increase in insulin sensitivity (principle of action 1): biguanides (metformin), thiazolidinediones (glitazones)
  • Increase in insulin release (principle of action 2): sulfonylureas, glinides
  • Inhibition of glucose absorption in the intestine (principle of action 3): acarbose, miglitol

In its guidelines mentioned above, the German Diabetes Society gives detailed suggestions for sequential differential therapy with these drugs (Fig. 1).
Based on this system, the most important data on the pharmacology and clinical application of these groups of substances are presented below (summarized in Tab. 1); In addition, an outlook on new antidiabetic drugs that are in approval or in clinical development or have just been approved is given.

Working principle 1 - increasing insulin sensitivity: biguanides (metformin)

The use of fenugreek preparations containing guanide (Galega officinalis) in the treatment of diabetes mellitus has been known since the Middle Ages [17]. After the lipophilic guanides phenformin and buformin were withdrawn from the market at the end of the 1970s due to the comparatively high risk of lactic acidosis, the more hydrophilic biguanide metformin (e.g. Glucophage®) is the only substance from this class still used in pharmacotherapy .

Metformin works by increasing cellular insulin sensitivity in the most important metabolically active tissues, i.e. liver, muscle and adipose tissue. The mechanisms discussed here are both a direct improvement in intracellular insulin signal transduction and the activation of an important energy sensor in the cell, namely the AMP-activated kinase [18]. This leads to increased mitochondrial oxidation of energy-rich substrates and thus to a consecutive increase in ATP production. Overall, the antidiabetic effect is based on the resulting decrease in hepatic glucose production and lipogenesis as well as a slight increase in glucose uptake in skeletal muscle. As a result of the increased insulin sensitivity, the body's own proinsulin and insulin levels decrease.

Clinical effects, effectiveness

With the mechanisms of action outlined above, metformin offers itself as the basis for the therapy of type 2 diabetes mellitus in overweight / obese patients. Clinically, after the start of metformin therapy, there is an improvement in the fasting glucose values ​​and at least an improvement in weight-reducing measures [19]. With metformin monotherapy in full doses, an improvement in the fasting glucose values ​​of around 35 to 70 mg / dl and a reduction in the HbA1c value by 1 to 2 percentage points can be expected after three months [20].

The long-term use of metformin was evaluated in the UKPD study (UKPDS, United Kingdom prospective diabetes study) [21]: During an average observation period of almost eleven years, a relative reduction in diabetes-related mortality of 42% compared to only dietetic patients can be reached; the corresponding number needed to treat (NNT, see box) was around 20. Furthermore, the incidence of diabetes-related macro- and microvascular complications decreased relatively by 32% (corresponding to an NNT over the entire observation period of around 10). In line with this, positive effects of metformin therapy on surrogate parameters of cardiovascular risk, such as C-reactive protein or plasminogen activator inhibitor-1, have been reported in various studies [22–24]. The effect of metformin on the lipid profile is slight: the LDL cholesterol value decreases somewhat, the HDL cholesterol value increases somewhat [25, 26].

Taken together, no other oral antidiabetic agent has similarly convincing endpoint data, so that metformin must continue to be the first drug therapy for overweight / obese type 2 diabetics.

Studies in which the clinical use of metformin is specifically examined in slim patients with type 2 diabetes mellitus have not yet been carried out. As far as can be deduced from individual subgroup analyzes and clinical-experimental investigations (e.g. [27]), a comparable effect of metformin on the blood sugar profile can in principle also be expected in this group.

Side effects and contraindications

Since the use of the lipophilic guanides phenformin and buformin was associated with an increased incidence of lactic acidosis, a fundamentally similar potential for side effects was postulated for metformin. However, therapy studies have so far shown no evidence of an increased rate of this potentially fatal complication with metformin [28, 29]. Lately, the risk of lactic acidosis has therefore been put into perspective with metformin [30]. However, the following contraindications must be observed before and during the use of metformin:

  • Renal insufficiency with a creatinine clearance of less than 60 ml / min: According to the recommendations of the German Diabetes Society, metformin is contraindicated if the serum creatinine value is over 1.2 mg / dl; However, due to the known limitations of this parameter, we recommend using creatinine clearance, for example calculated according to the so-called MDRD formula (MDRD = Modification of diet in renal disease study [31]).
  • All situations with the risk of central or peripheral hypoxia or substrate deficiency, e.g. B .:
    • Heart failure: The extent of the myocardial contraction restriction is not precisely defined here. Based on the idea of ​​an increased risk of lactic acidosis in peripheral hypoxia, we no longer use metformin from NYHA grade III heart failure.
    • Respiratory failure
    • Serious infections
    • Perioperative (from day -2)
    • Advanced malignancies
    • (planned) slimming cures
  • Disorders of the lactate metabolism:
    • Impaired liver function, as the liver plays a central role in lactate metabolism. The extent of the impairment of liver function at which a relevant risk of lactic acidosis begins is not known. For safety reasons, we discontinue metformin in all forms of liver cirrhosis.
    • Alcohol abuse, as ethanol itself inhibits the breakdown of lactate.
    • Planned intravenous administration of contrast medium (day -2 to +2)
  • pregnancy and breast feeding period

Other side effects, especially when the dose is increased too quickly, are gastrointestinal complaints such as bloating, inappetence, nausea, diarrhea or metallic taste, which can have a limiting effect on the patient's initial compliance [32].

After the contraindications have been ruled out, metformin therapy is generally started with a single dose of 500 mg in the evening. In order to minimize gastrointestinal side effects, the dose is increased every two weeks at least by 500 mg up to a final dose of 1,000 mg twice. An increase to three times 850 mg is possible, but probably does not bring any additional therapeutic benefit [20]. Regular diabetological follow-up checks - for example at three-monthly intervals - guarantee on the one hand the therapy efficiency by checking the blood sugar values ​​and the HbA1c value and on the other hand the therapy safety through the clinical and laboratory chemical exclusion of newly occurring contraindications. If the therapy is unsuccessful, a second oral antidiabetic is added according to the guidelines of the German Diabetes Society (Fig. 1). The UKPD study did, however, give indications of excess mortality in patients who were treated with metformin and sulfonylureas at the same time, so that this combination should be avoided, at least in patients with previous cardiovascular diseases [33].

Principle of action 1 - increase in insulin sensitivity: thiazolidinediones (glitazones)

The thiazolidinediones or glitazones were synthesized from the older group of fibrates as part of the development of a new lipid-lowering agent. The first clinical representative of the group, troglitazone, had to be withdrawn from the market in 2000 because of sometimes fatal hepatotoxic side effects [34–37]. Two products are now approved as successors, pioglitazone (Actos®) and rosiglitazone (Avandia®).

While fibrates act pharmacologically via the so-called PPAR-α (peroxisome-proliferator-activated receptor alpha), the thiazolidinediones activate the subform PPAR-γ. This pleiotropic protein, located in the cell nucleus, is mainly expressed in adipose tissue, but also in other tissues such as the pancreatic beta cells and, to a lesser extent, in the liver and muscles [38]. As an important transcriptional regulator of various metabolic processes in adipose tissue, it leads to increased fatty acid uptake and fat synthesis as well as increased glucose uptake. Adiponectin synthesis increases, 11-b-hydroxylase expression and thus local cortisol production decrease. Overall, there is increased adipogenesis in the area of ​​the subcutaneous adipose tissue, which probably reduces fat storage in other metabolically active organs and thus improves insulin sensitivity. This is shown in the liver by a reduced gluconeogenesis and in the muscle by an increased glucose uptake. In addition, PPAR-γ activation on the vascular endothelium, for example by reducing the expression of the inducible NO synthase, of adhesion molecules or of proinflammatory enzymes, appears to have an antiatherogenic effect [39].

Clinical effects, effectiveness

The described molecular effects result in various effects on the disturbed metabolism in type 2 diabetes mellitus. After several months of taking in the maximum dose (pioglitazone 30–45 mg, rosiglitazone 8 mg), pioglitazone and rosiglitazone reduce the HbA1c value by 1 to 1.5 percentage points . This applies both to monotherapy [40–42] and (additively) to combination therapies with metformin [43], sulfonylureas [44] or insulin [45], although the latter combination is not approved, at least in Germany [46, 47]. The risk of hypoglycaemia is extremely low with thiazolidinedione monotherapy [48].

Due to the fact that the thiazolidinediones have only been approved for a few years, long-term observations on the blood sugar-lowering effect are not yet available; however, the beneficial effects on glycemic control seem to persist for at least a few years [49]. In the recently published ADOPT study (A diabetes outcome progression trial) it was shown that in patients (n = 4,360) with untreated type 2 diabetes mellitus treatment failure occurred significantly less frequently with monotherapy with rosiglitazone than with monotherapy with metformin or gliburide [ 50].Therapy failure was defined as a confirmed fasting blood sugar level of> 180 mg / dl after at least six weeks of therapy with the maximum tolerated daily dose of the respective oral antidiabetic: it occurred after five years in 15% of the patients in the rosiglitazone group and in 21% and 32% of patients treated with metformin or gliburide (p <0.001, each for comparison with rosiglitazone). Furthermore, the insulin sensitivity was significantly increased compared to therapy with the two other oral antidiabetic agents and the loss of beta cell function was slowed down.

For pioglitazone, a minimal decrease in the LDL cholesterol serum value and a moderate reduction in the triglyceride serum value by 15 to 25% and an increase in the HDL cholesterol serum level by 10 to 15% have been described; In several studies, rosiglitazone led to a slight increase in the LDL cholesterol value of 10 to 15% [51]. The clinical relevance of these changes in the context of the above-mentioned postulated antiatherogenic effect of PPAR-γ activation is the subject of current investigations.

In the PROactive study (Prospective pioglitazone clinical trial in macrovascular events), it was examined in over 5,200 test persons with type 2 diabetes mellitus whether adding pioglitazone (30–45 mg / day) to existing antidiabetic therapy had an effect on the incidence of cardiovascular disease Had events [52]. In the combined primary endpoint (all-cause mortality, non-fatal myocardial infarction, stroke, acute coronary syndrome, vascular intervention / surgery, amputation) no significant effect of pioglitazone could be seen; a secondary endpoint (all-cause mortality, non-fatal myocardial infarction, stroke) occurred significantly less often (relative risk reduction: 16%). The NNT in this regard was given as 48 (over a period of three years). In the pioglitazone group, however, edema formation and non-fatal heart failure occurred more frequently, which at least significantly limits the clinical benefit from adding the substance.

The cardiovascular effects of rosiglitazone in combination therapy with other antidiabetic agents are currently being examined in the RECORD study (Rosiglitazone evaluated for cardiac outcomes and regulation of glycaemia in diabetes), which will run until 2009 [53].

In our opinion, based on the current study situation, the postulated cardiovascular benefit of a thiazolidinedione therapy cannot (yet) be validly assessed. In view of a blood sugar-lowering effect that is comparable to that of metformin, although the costs of these preparations are significantly higher, there is - contrary to other statements [54] - no reason to change the current therapy algorithms, which prevent the use of thiazolidinediones only in the second stage of oral antidiabetic therapy Suggest therapy [4].

Side effects and contraindications

After troglitazone was withdrawn from the market because of idiosyncratic hepatotoxicity with several cases of fatal liver failure, the occurrence of liver damage during therapy with pioglitazone or rosiglitazone has been carefully monitored [34–37]. A hepatotoxic class effect does not appear to be present, since during the use of pioglitazone or rosiglitazone an increase in transaminases over three times the upper norm only rarely (in 0.26 or 0.17% of the cases) and over ten times the norm only in individual cases has been observed [55, 56]. In the few published cases of severe liver damage, the hepatotoxicity was reversible after discontinuation of the respective preparation; So far, deaths cannot be clearly attributed to the currently approved thiazolidinediones [55, 57–59]. Another side effect of thiazolidinedione therapy is the weight gain already mentioned above (on average 2-4 kg), which is caused by both increased fluid retention and an enlargement of the subcutaneous adipose tissue. Cardiac decompensations can be indirectly triggered by this [60, 61]. Other - usually reversible - frequent side effects include changes in the blood count, paresthesia and gastrointestinal complaints. Accordingly, contraindications that emerge from the side effect profile are:

  • Pre-existing liver damage: This includes any significant, clinically or laboratory-chemically manifest liver dysfunction. An increase in alanine aminotransferase (ALT or GPT) over 2.5 times the upper limit of normal is an absolute contraindication. Whether this contraindication is limited in the future in obese patients with nonalcoholic hepatic steatosis who are likely to benefit from thiazolidinedione therapy is currently still open [62].
  • Any heart failure, this includes - unlike metformin - all NYHA grades I to IV.
  • Any insulin therapy: In some Anglo-Saxon countries, thiazolidinediones and insulin have been combined for a long time without a clinically relevant increased rate of side effects; In Germany, however, the combination is contraindicated because of a feared increase in the rate of heart failure [46, 47].

After the above-mentioned contraindications have been ruled out, thiazolidinedione therapy is started with 15 to 30 mg / day pioglitazone or 4 mg / day rosiglitazone and increased to 45 mg pioglitazone as a single dose or 4 mg rosiglitazone twice a day after eight weeks at the earliest. According to the specialist information sheet, the serum transaminases should be checked during therapy, "according to clinical judgment"; if the increase is more than three times the upper limit, a check is recommended as soon as possible and if the increase persists, immediate discontinuation of the thiazolidinedione preparation. Since the cases of severe hepatotoxicity described in the literature occurred two weeks to seven months after the start of thiazolidinedione intake, it seems to us that the frequency of controls is more frequent (e.g. clinical controls including liver values ​​and blood count after 1, 2, 4, 8 and 12 weeks , then quarterly in the further course). If there are clinical signs of liver dysfunction, an investigation must of course be carried out immediately.

Principle of action 2: Increase in insulin release

The group of insulin secretagogues is divided into the large subgroup of sulfonylureas and the smaller subgroup of glinides. The five most common representatives of the sulfonylureas are

  • Glibenclamide (e.g. Euglucon®),
  • Glibornuride (e.g. Gliborid®),
  • Gliclazid (Diamicron Uno®),
  • Glimepiride (e.g. Amaryl®) and
  • Gliquidon (Glurenorm®).

Tolbutamide, one of the first sulfonylureas, is rarely used today. The carbamoylbenzoic acid derivative nateglinide (Starlix®, which is still only approved in Germany in the form of a combination therapy with metformin) and the phenylalanine derivative repaglinide (NovoNorm®) are available for glinids.

Both the sulphonylureas and the glinides bind to (partly different) domains of the so-called sulphonylurea receptor SUR1 of the pancreatic beta cell, thereby closing the ATP-dependent potassium channels with subsequent depolarization of the beta cell. The subsequent opening of voltage-dependent calcium channels with an increase in cytosolic calcium leads to the indirect release of insulin [63, 64]. In a kinetic comparison, the binding of the glinides to the SUR1 receptor is faster and shorter than that of the sulfonylureas, which among other things is attributed to the lower risk of hypoglycaemia of the glinides compared to the sulfonylureas [65, 66]. The binding affinity of the individual substance class representatives to the sulfonylurea receptor of cardiac myocytes also varies considerably [67], which could be of considerable importance for potential cardiovascular side effects.

Clinical effects, effectiveness

The antihyperglycemic effectiveness depends on the beta cell reserve and in the initial stage of type 2 diabetes mellitus leads to a reduction in the HbA1c value of 1 to 2 percentage points for both sulfonylureas and glinides [33, 65, 68]. Since the UKPD study showed a benefit in terms of microvascular complications for diabetics with better glucose control, the macrovascular complications were not significantly reduced during therapy with sulfonylureas (in contrast to metformin), the indication for insulin secretagogues must be - especially as first-line therapy - should be viewed in a differentiated manner [33, 69].

Side effects and contraindications

The high protein binding of the sulfonylureas is important for the risk of hypoglycaemia as the most feared side effect. A significant increase in the effectiveness of the sulfonylureas must be expected, especially with concomitant medication with also high protein binding. The risk of hypoglycaemia exists particularly in the case of tight diabetics or irregular food intake and is exacerbated by alcohol consumption or renal insufficiency.

Since a body weight gain of 1 to 4 kg is to be expected within the first six months with sulfonylurea therapy, according to the current practice guideline of the German Diabetes Society, the sulfonylureas are only for normal weight (i.e. with a body mass index <25 kg / m2) therapeutics of first choice [4]. However, based on current retrospective studies on cardiovascular mortality in therapy with sulfonylureas, even this restricted recommendation remains worthy of discussion [70–72].

The question of the extent to which insulin secretagogues - possibly depending on their binding affinity to ATP-dependent potassium channels of the heart and blood vessels - have adverse effects on cardiovascular mortality and morbidity is still unclear even in the current literature ([73]) vs. [70, 72]). In addition to the disadvantages with regard to cardiac ischemia tolerance, direct arrhythmogenic effects in hypoglycemia or the toxicity of increased insulin levels could also play a role here. For a valid assessment of the clinical relevance of these pharmacodynamic parameters, prospective endpoint studies are required, as no reliable recommendation can be made on the basis of the retrospective observations available to date.

In any case, sulfonylurea therapy should be started with the lowest possible dose and, depending on blood sugar levels, only increased every two to four weeks. Since the sulfonylureas achieve more than 75% of their blood sugar-lowering effect at around 50% of the maximum dose, in our opinion the dose of a sulfonylurea should not be escalated to the maximum dose recommended by the manufacturer, but a combination therapy should be considered if the submaximal dose is insufficient (cave: metformin , see above).

With the glinids, too, should start with a low dose with each main meal (0.5 mg repaglinide or 60 mg nateglinide) and increase the dose slowly. In theory, glinides offer some advantages over sulfonylureas, but in clinical practice - similar to the debate about short-acting insulin analogues and conventional old insulin - these often appear less pronounced and are actually only insufficiently supported by long-term data.

Principle of action 3: inhibition of glucose absorption in the intestine

The representatives of this substance class approved in Germany are acarbose (Glucobay®, with significantly more extensive study evidence) and miglitol (Diastabol®).

The point of attack of the alpha-glucosidase inhibitors is the delay in the absorption of complex carbohydrates (no delay in the absorption of monosaccharides!) Due to the competitive binding to the alpha-glucosidase in the small intestine.

Clinical effects, effectiveness

The results of the STOP-NIDDM (Study to prevent NIDDM) have reassessed the alpha-glucosidase inhibitors due to the delay in progression shown here from pathological glucose tolerance to manifest diabetes mellitus [74, 75]. This diabetes prevention study also showed that acarbose significantly reduced the risk of cardiovascular disease and arterial hypertension, although it should be noted that lifestyle changes (diet, physical activity) proved to be even more effective than any drug treatment. The STOP-NIDDM, however, was not concerned with the treatment but rather the prophylaxis of type 2 diabetes mellitus.

A current meta-analysis based on 41 individual studies exists on the therapeutic use of alpha-glucosidase inhibitors in existing type 2 diabetes mellitus [76]. Accordingly, alpha-glucosidase inhibitors reduce the postprandial glucose increase by around 50 mg / dl, but the fasting blood sugar level also drops by around 20 mg / dl. Compared to placebo, acarbose results in a non-dose-dependent reduction in the HbA1c value of 0.8 percentage points and with miglitol a rather dose-dependent reduction in the HbA1c value of 0.7 percentage points, whereby the antihyperglycaemic effect of the alpha-glucosidase inhibitors is greater than that of the other oral ones Antidiabetic drugs appear inferior. The meta-analytical evaluation did not show any clinically relevant changes for the lipid profile - in particular, the positive effect on the triglyceride level postulated in some individual studies could not be confirmed. The body weight of diabetics remains unchanged after therapy with alpha-glucosidase inhibitors. Furthermore, an influence on mortality or diabetes-specific secondary complications with alpha-glucosidase inhibitor therapy cannot be proven meta-analytically, although the duration of the studies included in the meta-analysis was rarely more than a year [76].

Side effects and contraindications

The gastrointestinal side effects in the form of nausea, flatulence, abdominal pain and diarrhea are frequent and limit its use. They mainly occur when patients consume too few complex carbohydrates in their diet.

Since the side effects show a strong dose-dependency, it is advisable to gradually dose the alpha-glucosidase inhibitors (e.g. start with 25 mg acarbose in the evening, increase by 25–50 mg / week up to a maximum dose of 3 times 100 mg / Day).

New approaches in oral antidiabetic therapy

In recent years, based on new findings on the physiology of beta cell secretion and insulin resistance, some interesting therapeutic approaches for type 2 diabetes mellitus have been developed (Tab. 2). In view of the short test phase, endpoint data on diabetes-associated complications, cardiovascular diseases or mortality, which would be necessary for a comprehensive assessment, are still largely pending for the substances mentioned. New developments in drug therapy for obesity, such as rimonabant (Acomplia®), can have a positive effect on glucose metabolism, cannot be discussed here.

Oral glucose supply stimulates the release of so-called incretins in the intestinal mucosa. Irrespective of the serum glucose level, these peptide hormones cause beta-cellular insulin release, which in vivo is probably responsible for 50 to 70% of the postprandial increase in insulin [77]. This mechanism is used for various pharmacological approaches.

A synthetic analogue of exendin-4, a protein from the saliva of the North American Gila monster (Fig. 2), which is structurally similar to the human incretin GLP-1 (glucagon-like-peptide 1) and acts on the human GLP-1 receptor , was approved in the USA in spring 2005 for combination therapy with metformin or sulfonylureas for poorly controlled type 2 diabetes mellitus; The Europe-wide approval took place in November 2006. It has been available in Germany since April 1, 2007.

According to the data available so far, exenatide reduces the dose of 10 µg / day s.c. the HbA1c value on average by just under 1 percentage point; In addition, most patients lose a few kilograms in weight [78]. So far, gastrointestinal complaints such as nausea have mainly been mentioned as side effects; the risk of hypoglycaemia is low. Long-term data are not yet available.

Synthetic GLP-1 analogs are currently being developed. Most of the data on this are currently available for liraglutide. Liraglutide, which also has to be administered subcutaneously, appears to have an effect / side effect profile similar to exenatide [54] and is being further tested in phase III studies.

Inhibitors of the enzyme dipeptidyl peptidase IV, which normally breaks down GLP-1 in the blood within a few minutes, are in clinical trials. For example, the orally administered vildagliptin (Fig. 3) in combination with metformin was able to lower the HbA1c value by 0.5 percentage points and improve the beta-cellular insulin response in a 52-week study; essentially there were only slight gastrointestinal side effects [79]. Sitagliptin (Fig. 3) and saxagliptin follow the same principle of action and appear to improve glucose tolerance in short-term use [80]; Observational studies over a longer period have not yet been published here.

Like insulin, amylin is secreted by the beta cells and lowers the postprandial blood sugar increase by inhibiting glucagon secretion and gastric emptying. In addition, it has an appetite-reducing effect [81]. The amylin analog pramlintide was approved in the USA at the beginning of 2005 for patients who were poorly controlled on intensive insulin therapy. The preparation must be administered subcutaneously with each meal. Clinical data show a reduction in postprandial hyperglycaemia by up to 80% through the use of pramlintide in combination with insulin lispro [82]. Data on prolonged use show a reduction in the HbA1c value of 0.6 percentage points after 52 weeks, the weight decreased by an average of 1.6 kg. The main side effect is nausea, which in individual studies occurred in almost a quarter of the test subjects [83]; The mechanism of action also results in the risk of an increased tendency to hypoglycaemia. The manufacturer has not provided any information on approval in Europe.

The use of PPAR-α / γ agonists can theoretically be justified due to their double mechanism of action and is successful in animal experiments. However, the substances available to date do not seem to be safe to use in humans: The development of muraglitazar (Fig. 4) was discontinued because a higher incidence of cardiovascular events and heart failure was noticed during clinical trials of the preparation [84]. Tesaglitazar is not being developed further as there was evidence of nephrotoxicity. At the moment it remains to be seen whether new developments with this approach will actually find their way into practical clinical application.

Importance of insulin therapy and time of start

Insulin requirement for type 2 diabetes mellitus is not the exception, but sooner or later the rule, provided that the glycemic therapy goals are to be adhered to. Already at the time of the initial diagnosis there is not only an increased insulin resistance, but also the beta-cellular insulin capacity has already fallen by half compared to a healthy normal collective [85]. The beta cell failure progresses in the further course, so that after about six years of the disease about half of all patients need insulin for glycemic control [86]. Oral antidiabetic drugs are therefore sooner or later limited in their blood sugar-lowering effectiveness: on the one hand, increasing the sensitivity of insulin with metformin or thiazolidinedione only affects the aspect of insulin resistance and excludes progressive beta cell failure as an equally important pathogenetic factor [81] Eight. On the other hand, substances that mobilize the insulin reserves of the pancreatic beta cells, such as sulfonylureas or glinides, are only effective as long as there is a natural insulin reserve. Related to this is the now outdated term of secondary failure, which is described as a therapeutic failure of insulin secretagogues despite optimal lifestyle circumstances, i.e. those that minimize peripheral insulin resistance (such as normal weight, adherence to dietary requirements, regular physical activity). is defined [87, 88].

The national health care guidelines under the auspices of the German Medical Association (NVL T2DM 2002, [10]) and the current practice guideline of the German Diabetes Society from 2006 [4] recommend starting insulin therapy for type 2 diabetes mellitus, if despite the use of two oral antidiabetic drugs, the HbA1c value is permanently (i.e. over longer than three months) over 7% (Fig. 1). Similarly, the guidelines of the International Diabetes Federation (IDF) from 2005 recommend starting insulin therapy if the HbA1c value exceeds 7.5% despite optimization of oral antidiabetic therapy and individual exhaustion of lifestyle modification [12] .

In practice, the doctor's recommendation to start insulin therapy often provokes a defensive attitude on the part of the patient, which in up to half of all patients is so pronounced that the further success of the treatment is endangered [89]. The main cause is the change in the individual experience of illness associated with the therapy change ("needle dependency" with fear of the invasiveness of the treatment, social stigmatization and excessive demands), which is additionally reinforced by traditional ideas about pain, side effects and chances of success of insulin therapy [90]. In order to break such psychological insulin resistance [91], it is necessary - even more than with oral diabetes therapy - to continuously educate and care for the patient in the sense of patient education with the central goal of self-care [92, 93].

Therapy options for insulin therapy

Insulin therapy for type 2 diabetes mellitus differs from that of type 1 diabetes mellitus, among other things, in that it does not necessarily have to be carried out in the form of conventional (CT) or intensified conventional (ICT) monotherapy, but rather between pure oral therapy Antidiabetic drugs and pure insulin therapy a combination therapy with insulin and oral antidiabetic drugs can be interposed or maintained permanently [94–96].

Basal-assisted oral anti-diabetes therapy

The most frequently used variant of a combination therapy with insulin is to continue the therapy with oral antidiabetic drugs with the additional administration of a delayed insulin. This treatment strategy, also known as basal-assisted oral anti-diabetes therapy (BOT) in the German-speaking world, was subjected to critical meta-analyzes 15 years ago [97, 98] and was then initially forgotten. The combination of oral antidiabetic agents with delayed insulin only received real attention in the last five years with the introduction of the long-acting analog insulin insulin glargine (Lantus®) [96, 99, 100]. In the studies and analyzes published on this, the focus was initially less on the effectiveness of the therapy, but primarily on the safety aspects of the insulin glargine, which is applied once at night, which then also affects the hypoglycemia frequency of the single administration of a conventional neutral protamine hawthorn delay insulin (NPH insulin). appeared superior [99, 100].

Clinical effects, effectiveness

On the therapeutic effectiveness of a BOT, half-yearly studies show that the above-mentioned therapeutic goals can be achieved in principle for both conventional NPH insulin and insulin glargine. In the so-called treat-to-target study [99], fasting blood sugar levels of around 120 mg / dl and average HbA1c values ​​of 7% were achieved. In a study from 2005, which compared a BOT with insulin glargine to conventional insulin therapy, the subjects treated with a BOT achieved fasting blood sugar values ​​averaging 115 mg / dl; the HbA1c value was below 7% in half of the patients (HbA1c mean value of the entire BOT collective 7.15% ± 0.90%) [101]. In general, instead of administration of insulin glargine or NPH insulin, administration of the long-acting analog insulin insulin detemir (Levemir®) is conceivable, although the literature on this is sparse [102].

After checking the approval compatibility of oral antidiabetic drugs and insulin, a BOT with ten units of NPH insulin or insulin glargine can be started at night according to the literature [99, 101]. The further adjustment of the insulin dose should then be carried out weekly based on the capillary measured fasting blood sugar level and a target value of 100 mg / dl according to the following scheme:

  • With fasting blood sugar> 100 mg / dl and <120 mg / dl ⇒ +2 units of insulin
  • With fasting blood sugar ≥ 120 mg / dl and <140 mg / dl ⇒ +4 units of insulin
  • With fasting blood sugar ≥ 140 mg / dl and <180 mg / dl ⇒ +6 units of insulin
  • With fasting blood sugar ≥ 180 mg / dl ⇒ +8 units of insulin (also consider or exclude nocturnal hypoglycaemia!)

According to the literature, three to five hypoglycemia levels below 60 mg / dl per patient per year must be expected with such a therapy regimen, although severe hypoglycaemia, i.e. those requiring external help, is the absolute exception [99, 101].

Therapy with mixed insulins / conventional insulin therapy

If the basal-supported oral anti-diabetes therapy does not succeed in achieving normal values ​​for fasting blood sugar, HbA1c and postprandial blood sugar, instead of delayed insulin, an attempt can also be made with an evening preprandial administration of a mixed insulin, i.e. a short-acting insulin and NPH insulin in one A fixed mixing ratio of, for example, 30% to 70% [103], while the oral antidiabetic concomitant medication is continued [104]. In the so-called 1-2-3 study, whose mixed insulin was based on the insulin analogue insulin aspart (NovoRapid®, NovoMix® = In-sulinaspart, biphasic), the quantitative retention of the previous dose is recommended as the starting dose, i.e. instead of the delayed insulin now the same amount of mixed insulin. Only with a previous dose of delay insulin of ≥ 30 units should an additional initial increase of the mixed insulin amount by 30% occur [104]. The further adjustment of the insulin dose then takes place every three to four days based on the capillary measured fasting blood sugar level with a target range between 80 to 110 mg / dl:

  • With fasting blood sugar ≤ 80 mg / dl ⇒ -3 units of insulin
  • With fasting blood sugar> 80 mg / dl and ≤ 110 mg / dl ⇒ no change
  • With fasting blood sugar> 110 mg / dl and ≤ 140 mg / dl ⇒ +3 units of insulin
  • With fasting blood sugar> 140 mg / dl and ≤ 180 mg / dl ⇒ +6 units of insulin
  • With fasting blood sugar> 180 mg / dl ⇒ +9 units of insulin

Such a regimen can reduce the HbA1c value to below 6.5% in 20% of all patients, with around 15 mild hypoglycaemia episodes per patient year - but hardly any more severe hypoglycaemia [104].

At the latest if this therapy regimen also fails (e.g. failure to achieve an HbA1c value of less than 7% after 16 weeks [104]), the oral antidiabetic concomitant medication (with the possible exception of metformin [105]) should be discontinued and in addition to the evening mixed insulin administration also a mixed insulin for breakfast (starting dose 3 units if the fasting blood sugar value ≤ 110 mg / dl, or 6 units if the fasting blood sugar value> 110 mg / dl). The further adjustment of the morning mixed insulin administration takes place on the basis of the blood sugar value measured before dinner in analogy to the algorithm given above. This double or even triple (here an additional 3 units of mixed insulin before lunch and further adjustment for late evening blood sugar values> 140 mg / dl [104]) administration of mixed insulin appears to be a therapy consisting of oral antidiabetic and delayed insulin or oral antidiabetic and once a day To be superior to mixed insulin [106, 107], whereby in 40 to 60% an adequate glycemic control corresponding to an HbA1c value of below 6.5% can be expected [104, 108].

In the 1-2-3 study mentioned above, a mixed insulin based on the analogue insulin aspart was used [104]. Various studies have dealt with the question of whether the more expensive mixed insulins based on analog insulin [103] actually show advantages over the conventional, cheaper mixed insulins. In terms of glycemic control and hypoglycemia frequency, however, the differences appear rather marginal [109–112].

Supplementary therapy with short-acting insulins or inhalable insulin

A possible alternative is the so-called supplementary insulin therapy, or SIT for short. This is the preprandial, meal and blood sugar-adapted administration of short-acting insulin with [113] or without [114, 115] additional oral antidiabetic agents. In this context, mention should also be made of the recently introduced possibility of inhaling the short-acting insulin in the form of Exubera®, the only substance that has been approved so far. An additional therapeutic benefit of the significantly more expensive Exubera® compared to short-acting human insulin or short-acting analog insulins is not seen by the Institute for Quality and Scientific Research in Health Care (IQWiG) according to a statement from 2006 [116].

In general, the data situation on SIT is thin. The results of one of the few studies that have recently dealt with this form of insulin therapy indicate inadequate glycemic control with an average HbA1c value of 7.4% ± 1.7% after 24 ± 14 -month treatment duration. A further complicating factor is that in this study, almost half of the patients (42%) still required the additional administration of late-night insulin [114]. In a six-month study based on the analog insulin Insulinlispro (Humalog®), the form of therapy for SIT was compared with three administration of a corresponding mixed insulin, a lower percentage of HbA1c values ​​below 7% is described for SIT (40 , 4% with SIT vs. 59.3% with mixed insulin) [115]. The combination of SIT with metformin seems somewhat more effective: Here, in the PHAZIT study (PHAZIT = pharmaco-economic application observation for supplementary insulin therapy) after six months of treatment, previously poorly controlled type 2 diabetics (initial HbA1c 8.8% ± 1 , 1%) HbA1c values ​​of 7.2% ± 1.0% (normal insulin plus metformin) and of 7.1% ± 1.1% (insulin aspart plus metformin) were achieved [113].

In a six-month study with type 2 diabetics who were inadequately controlled with metformin, the addition of Exubera® resulted in a somewhat greater drop in the HbA1c value of 2.03% on average [117].

Intensified insulin therapy

If sufficient metabolic control cannot be achieved with the previously mentioned forms of insulin therapy, and if the patient is sufficiently compliant, intensified conventional insulin therapy according to the basic bolus principle (ICT) should be initiated. In general, it must be noted that ICT in type 2 diabetics (also referred to in German-speaking countries as complementary intensified insulin therapy [KIT]) is less standardized than for type 1 diabetics [117a]. Since the ICT or KIT comes closest to physiological insulin secretion [118], it is considered the gold standard for insulin administration. In the DCC-Trial / EDIC-Study (DCCT / EDIC = Diabetes control and complications trial / Epidemiology of diabetes interventions and complications) - but only for type 1 diabetics - the best evidence for the general superiority of ICT over conventional Administration of mixed insulins both in glycemic control and in avoiding long-term vascular damage [119, 120]. This fact can also be adopted analogously for type 2 diabetics [120a], although there are also direct indications for this, namely in the context of the Kumamoto study (study was limited to normal weight type 2 diabetics [121, 122]) and - with methodological restrictions - also within the scope of the UKPD study [123]. In contrast, a study at national level found no relevant advantage of ICT over conventional mixed insulin therapy [124].

In general, the ICT places very high demands on the cooperation of the patient, because reliable self-monitoring of blood glucose (SMBG = self-monitoring of blood glucose) is indispensable before the regular, food-dependent insulin administration. The minimum is the four daily measurements before each main meal and before going to bed. A common therapeutic concept for ICT in patients with type 2 diabetes mellitus is to start with a total insulin dose of 0.3 units per kg of body weight, with 50 to 70% of the total dose being divided into the three main meals in the form of an old insulin or short-acting analog insulin and the remainder of the total dose is administered in the form of a human delayed insulin or long-acting analog insulin. The further adjustment is then made on the basis of the SMBG, with deviations in the preprandial or fasting blood sugar values ​​being corrected successively by 2 to 4 units by adjusting the next preceding insulin application [125]. Despite all the advantages that ICT brings for a self-responsible patient, it should not be overlooked that especially with ICT of type 2 diabetics for optimal glycemic control it is not uncommon for daily total insulin doses of more than 100 units, i.e. about 1 Unit per kg of body weight, and drastic weight gains of up to 10 kg are possible [126]. In individual cases, additional oral diabetes therapy can also be continued under ICT, as long as - according to an empirical rule of thumb - at least 10 units of insulin can be saved. Maintaining metformin can also have beneficial effects on weight gain during insulin therapy.

Importance of the analog insulins

In addition to the previously presented types of insulin therapy options with

  • basal-assisted oral anti-diabetes therapy (BOT),
  • one to three times administration of a mixed insulin in the sense of a conventional insulin therapy (CT),
  • sole or supplementary meal-related administration of normal insulin or inhaled insulin (SIT) and
  • Insulin therapy according to the basic bolus principle in the sense of an intensified conventional insulin therapy (ICT)

In terms of differential therapy, the question arises of the extent to which either conventional old or delayed insulins or short-acting or long-acting analog insulins are used for this purpose. For the pharmacological as well as economic characterization of these types of insulin, reference is made to two review articles that were recently published in this journal [103, 118]. In this context, a resolution passed by the Federal Joint Committee (G-BA) on July 18, 2006 is particularly relevant for Germany, which stipulates that the short-acting insulin analogues in type 2 diabetics will generally only be charged to statutory health insurances in the future (GKV) may be prescribed if they are not more expensive than conventional human insulins [127]. The background to this is a critical statement written by IQWiG last year on the benefits of insulin lispro and insulin glulisine (Apidra®) compared to conventional human insulin [128]. The assessment is based on study reports from the relevant manufacturers as well as on five comparative studies from the literature [8, 129–132] and refers to larger meta-analyzes from 2004 and 2005 [133, 134]. Based on this primary literature, the IQWiG's conclusion on the non-superiority of the short-acting analog insulins appears to be consistent, but it can hardly do justice to the inter-individual, difficult-to-standardize heterogeneity of diabetes therapy and the methodologically insufficiently recorded parameters such as quality of life and patient satisfaction. For doctor and patient, it is generally desirable that the entire differential therapeutic armamentarium for diabetes therapy is still available - this would be easily possible in accordance with the decision of the G-BA with a price adjustment of the short-acting analog insulins on the part of the manufacturer.

The therapy of type 2 diabetes mellitus nowadays has ambitious goals which, in addition to a change in the patient's lifestyle, in most cases require oral antidiabetic therapy and / or insulin therapy to be achieved. In oral antidiabetic therapy, metformin is the starting point for the generally overweight type 2 diabetic; It is important to observe the specified contraindications, which may have to be redefined in the next few years. Oral combination therapy is possible with a large number of preparations; The decision in favor of the metformin combination partner must be made individually, taking into account the daily blood sugar profile (e.g. high postprandial values ​​ combination with glinide or acarbose, persistently high fasting blood sugar value  combination with a thiazolidinedione) and the comorbidities.

For the sulfonylureas and glinides, prospective long-term studies are urgently required, which would allow a more precise assessment of cardiovascular side effects beyond the antihyperglycaemic effectiveness of the individual insulin secretagogues.

In the next few years, new drugs with theoretically convincing principles of action will be approved; To what extent this translates into tangible benefits for the patient remains to be seen.

In spite of the broad differential therapeutic spectrum of oral antidiabetic agents, the administration of insulin for adequate glycemic control is unavoidable in about half of all patients with type 2 diabetes mellitus after a short period of time. In addition to the classic forms of insulin monotherapy (conventional insulin therapy or intensified conventional insulin therapy), a combination therapy consisting of oral antidiabetic agents and delayed and / or short-acting insulin can initially be successful. The current state of studies on the efficiency and effectiveness of the individual forms of insulin therapy in type 2 diabetes mellitus (with the exception of the more intensive conventional insulin therapy) is still confusing, so that clear recommendations for differential therapy cannot be given.

After completion of the manuscript, data from the ADOPT study became known that indicated an approximately 2-fold increased risk of fractures for women on rosiglitazone therapy. A current warning from the American regulatory authority (FDA) underlines that a class effect of the thiazolidinediones is likely to be assumed. Corresponding red-hand letters from the manufacturers Takeda and Glaxo-Smith-Kline were sent in March 2007.

The bibliography as PDF

Priv.-Doz. Dr. med. Cornelius Bollheimer, Clinic and Polyclinic for Internal Medicine I
(Director: Prof. Dr. J. Schölmerich), University of Regensburg, 93042 Regensburg, email: [email protected]
Dr. med. Christiane Girlich, Dr. med. Ulrike Woenckhaus, Priv.-Doz. Dr. med. Roland Büttner, Clinic and Polyclinic for Internal Medicine I, University of Regensburg, 93042 Regensburg

Fig. 1. Therapy algorithm for diabetes mellitus type 2. The presentation essentially follows the current DDG therapy guideline [4], although this does not prescribe a three-month observation phase for basal supported oral therapy before the start of conventional or intensified insulin therapy.
Caution: The schematic procedure can only be recommended for clinically stable patients with moderate hyperglycaemia; insulin therapy may be necessary at any level in the event of derailments.
* A combination of sulfonylureas and metformin is not recommended
BMI = body mass index

Tab. 1. Oral anti-diabetic drugs (the dose recommendations correspond to the information in the respective specialist information)

Oral anti-diabetic drug

Standard dosage

Dosage in case of renal insufficiency

Combination with insulin

Metformin (e.g. Glucophage®)

2 x 850-1,000 mg / d

Administration up to creatinine clearance> 60 ml / min possible,
Contraindicated from creatinine clearance <60 ml / min


Glibenclamide (e.g. Euglucon®)
Glibornuride (e.g. Gluborid®)
Gliclazid (e.g. Diamicron Uno®)
Glimepiride (e.g. Amaryl®)
Gliquidon (Glurenorm®)

1-3 x 1.75-3.5 mg / d
1-3 x 12.5-25 mg / d
1 x 30-120 mg / d
1 x 1-6 mg / d
1 x 15-120 mg / d

If necessary, dose reduction if creatinine clearance <60 ml / min, discontinue if creatinine clearance <30 ml / min

Up to creatinine clearance> 30 ml / min, no change,
then nominally contraindicated

Yes (basal insulin)

Pioglitazone (e.g. Actos®)

Rosiglitazone (e.g. Avandia®)

1 x 15-45 mg / d

1 x 4-8 mg / d

No dose adjustment necessary up to creatinine clearance> 15 ml / min, contraindicated if dialysis is required
No dose adjustment is necessary up to creatinine clearance <30 ml / min, only with caution in the case of severe renal insufficiency


Nateglinide (e.g. Starlix®)

Repaglinide (e.g. NovoNorm®)

3 x 60-180 mg / d

3 x 0.5-4 mg / d

From creatinine clearance <50 ml / min dose adjustment according to blood sugar values, in principle also possible for dialysis patients

Yes (basal insulin)

Alpha-glucoside inhibitors:
Acarbose (e.g. Glucobay®)
Miglitol (e.g. Diastabol®)

1-3 x 50-100 mg
1-3 x 50-100 mg

No dose adjustment until creatinine clearance ≥ 25 ml / min; contraindicated for creatinine clearance <25 ml / min


Number needed to treat (NNT)

The NNT is the reciprocal value of the absolute risk reduction; it indicates how many patients have to be treated in order to achieve a defined treatment goal for one patient.
The absolute risk reduction due to the treatment is the risk difference between the placebo and verum groups.

Tab. 2. Principle of action of new antidiabetic drugs [mod. Representation according to 135]




Increase in insulin secretion



Inhibition of glucagon secretion




Inhibition of gastric emptying



Increase in satiety, weight loss



GLP-1 = glucagon-like-peptide 1; DPP IV = dipeptidyl peptidase IV

Fig. 2. The North American crustacean Heloderma suspectum; their saliva contains exendin-4, the model for the synthetic incretin mimetic exenatide [Photo: Lilly / Roetz]

Fig. 3. Vildagliptin (Novartis) and sitagliptin (MSD Sharp & Dohme) - orally administered inhibitors of dipeptidyl peptidase IV, which are in phase III of the clinical trial (sitagliptin was approved in Europe at the end of March 2007 [Januvia®; combination with metformin or Thiazolidinedione, if the administration of metformin or a thiazolidinedione plus lifestyle change for blood sugar control was not sufficient], in Mexico and the USA it was already approved [Januvia®; monotherapy or combination with metformin or thiazolidinedione] in patients with a change in lifestyle, insufficient blood sugar control could be achieved)

Fig. 4. Muraglitazar and Tesaglitazar: Dual-acting PPAR agonists are theoretically particularly interesting new substances for the therapy of type 2 diabetes mellitus due to their double mechanism of action, which in particular reduces insulin resistance and additionally improves parameters of lipid metabolism; However, the development of both compounds is currently not being pursued further because toxicity problems arose (cardiotoxicity, nephrotoxicity)