Heiner Laube, Department of Internal Medicine, Universität Giessen, Giessen, Germany

Clin Drug Invest. 2002;22(3) 

In This Article

Structure, Mode of Action and Pharmacological Properties

Acarbose is a pseudotetrasaccharide, a natural microbial product derived from culture broths of Actinoplanes strain SE 50. The unsaturated cyclitol component of the molecule has been identified as essential for a-glucosidase inhibitory activity.[5]

Acarbose binds reversibly, competitively and in a dose-dependent manner to the oligosaccharide binding site of a-glucosidase enzymes in the brush border of the small intestinal mucosa. As a consequence, hydrolysis is prevented. This effect lasts for 4 to 6 hours provided that acarbose is present at the site of enzymatic action at the same time as the oligosaccharides. Thus acarbose must be administered with the first bite of a main meal.

Acarbose binds to intestinal sucrase with a 104 - to 105 - fold greater affinity than sucrose. The drug delays the intestinal hydrolysis of oligo-and disaccharides by a-glucosidases, mainly in the upper half of the small intestine. Consequently, the absorption of monosaccharides after a meal is delayed and transport through the mucosal surfaces into the circulation is interrupted. The suppression of a-glucosidase is reversible, although pharmacological activity is reliable and persistent with long-term use. Effects with continued use can be maintained over years[6,7] and no reportsof acarbose failure are present in the available literature. There is no need for dosage adjustment in slight renal insufficiency; however, acarbose should be withdrawn in case of severe progressive renal insufficiency.

Hara et al.[8] suggested that to obtain the greatest benefit from acarbose, strict adherence to a high carbohydrate (>50% of energy intake) diet is required. More recent findings have contradicted this theory.[9,10]

The relative affinity of acarbose for specific enzymes is as follows: glycoamylase > sucrase > maltase > dextranase. Acarbose has little affinity for isomaltase and no affinity for the ß-glucosidases, such as lactase. Recent evidence also suggests that acarbose inhibits pancreatic a-amylase.[11] Acarbose has no direct effect on the absorption of glucose.

With long-term acarbose administration, glucosidase activity increases slowly in the lower half of the small bowel.[12] As the distal part of the small intestine is less exposed to complex carbohydrates, a-glucosidase activity is relatively low, especially with a northern European diet. This is even more pronounced with a 'fast food' diet with high glycaemic index and little dietary fibre.

a-Glucosidase activity in the small intestine is subject to interindividual and racial differences.[13,14] Initial therapy with an a-glucosidase inhibitor often results in carbohydrates appearing in the colon, where bacterial fermentation may occur, accounting for the frequency and severity of gastrointestinal adverse effects, predominantly flatulence and loose stools. The quantity of undigested carbohydrates reaching the colon can be determined by analysis of breath hydrogen. Acarbose should be started at a low dose followed by slow upward dose titration to reduce or avoid gastrointestinal adverse effects.[15]

Long-term treatment with acarbose increases colonic bacterial mass, that of lactobacteria in particular.[16,17] This can impair carbohydrate absorption and cause increased bacterial carbohydrate fermentation and faecal acidification. In patients with liver cirrhosis and portosystemic encephalopathy, this action of acarbose partially mimics the effect of lactulose or lactitol.[18] Acarbose has a favourable therapeutic profile for the long-term treatment of patients with type 2 diabetes and liver cirrhosis. The drug has been reported to lower levels of ß-hydroxybutyrate significantly (p < 0.01), and to reduce hyperammonaemia and gut pH.[19]

Acarbose treatment may be associated with a reduction in the risk of colon carcinogenesis. Diabetes itself is associated with a 30% increase in colon cancer risk.[20] Holt et al.[20] suggest that the observed changes in bacterial flora, decreased stool pH and ß-hydroxybutyrate seen with acarbose treatment are associated with anti-proliferative effects in colonic epithelial cells that may potentially decrease the risk of carcino-genesis. Definite evidence of risk alteration has not yet been established; any inherent antiproliferative benefits of acarbose are unlikely to emerge for at least 10 years.[20,21]

Acarbose is poorly absorbed and systemic bioavailability is low. After oral administration, <2% of the unchanged drug is absorbed and enters the circulation, with most remaining in the lumen of the gastrointestinal tract. Acarbose is cleaved in the large intestine by bacterial enzymes into several metabolisable and absorbable intermediates (glucose, maltose, acarviosine), approximately 35% of which will be absorbed, depending on the microbial flora in the intestine. The absorbed material appears in the urine as metabolites, mostly glucose, within 14 to 24 hours.[22] Excretion via the kidneys predominates. The small quantity of drug that reaches the systemic circulation may enter the islet lysosomal-vacuolar apparatus by endocytosis and can induce a slight but insignificant suppression of glucose-and sulphonylurea-stimulated insulin release.[23,24]

After a high carbohydrate meal, acarbose lowers the postprandial rise in blood glucose by approximately 20%, or 2.75 to 3.30 mmol/L, depending on the dose, the extent of hyperglycaemia and the type of carbohydrate ingested.[25] There is a greater effect on postprandial hyperglycaemia after ingestion of starch than of sucrose.[26] This effect is additive with that of dietary manipulation[27] and ismore pronounced in patients with newly diagnosed diabetes.[28]

Acarbose significantly lowers postprandial blood glucose measured 60, 90 and 120 minutes after a meal.[29] The effects can be seen after the first dose and can last for 3 to 5 hours, although an acute effect is apparent within a few minutes.[30] The decreased peak plasma glucose and increased time to plasma glucose nadir collectively inhibit reactive hypoglycaemia by decreasing the hyperglycaemic stimulus to insulin secretion.

Several clinical trials confirm that acarbose monotherapy lowers fasting blood glucose significantly (p < 0.02) by 1.1 to 1.3 mmol/L, or approximately 10 to 15%, in patients with type 2 diabetes.[27,30,31,32,33,34] The mechanism for this effect is secondary to the lowering of postprandial hyperglycaemia. A reduction in glucose toxicity, through decreased postprandial hyperglycaemia and lower plasma triglycerides, has also been reported following acarbose treatment.[35]

Decreased insulin resistance following acarbose treatment occurs in patients with impaired glucose tolerance and confirmed type 2 diabetes. In hypoglycaemic clamp studies there was a significant difference (p < 0.05) in the change of insulin sensitivity in 45 patients in response to acarbose compared with placebo.[34] The beneficial effect of acarbose on fasting plasma glucose is probably potentiated by an increased late rise in glucagon-like peptide 1 (GLP-1).[35] The initial improvement in blood glucose with acarbose tends to be modest, but with long-term use efficacy steadily improves. Benefits are more pronounced after 3 months and are maintained over several years without evidence of decreased effect or treatment failure.[27] The UK Prospective Diabetes Study (UKPDS), for example, demonstrated that in 309 patients with type 2 diabetes the hypoglycaemic and HbA1c-lowering effect of acarbose, when used in combination therapy, was highly significant (p < 0.001) and sustained for at least 3 years.[36]

The reduction in blood glucose concentration following acarbose treatment is accompanied by decreased plasma insulin, both a significant lowering of fasting insulin and a reduction in the post-prandial insulin rise.[37] Postprandial insulin levels increased significantly less (p= 0.02) in 72 type 2 diabetic patients treated with acarbose compared with gliclazide over a total period of 72 weeks.[38] In elderly patients with type 2 diabetes, this reduction in insulin has been shown to be dose dependent and can reach 60%.[34,39,40] The decrease in insulin secretion is secondary to reduced postprandial glucose and is most dramatic in individuals with high insulin secretory rates. No change[41,42] or an increase[43] in insulin secretion may occur in patients with inherently low insulin secretion. No effects of acarbose on first-or second-phase insulin release during a hyperglycaemic clamp are evident.

There is no detectable effect of acarbose on energy or nutrient intake, and patients' eating habits are not markedly changed.[26,44] Although animal studies have repeatedly shown consistent bodyweight loss following acarbose, the majority of studies in lean or obese patients with type 2 diabetes have shown no[45,46,47] or minimal[9] body-weight loss. The energy lost in the stool due to carbohydrate malabsorption is unlikely to be significant. In contrast, patients on insulin or glibenclamide (glyburide) therapy are predisposed to bodyweight gain, typically up to 3 to 5kg.[3]

This difference in action can be explained by the fact that acarbose decreases postprandial insulin and affects the brain's satiety centre by increasing the level of GLP-1,[48] whereas long-term treatment with glibenclamide and insulin infusion significantly increases leptin concentration (p < 0.01), which is responsible for obesity. Obese nondiabetic humans have increased leptin levels and may be resistant to any regulatory effects of leptin on bodyweight. Acarbose did not induce changes in leptin in a 16-week clinical trial in patients with type 2 diabetes, but does lead to a fall of post-prandial insulin concentrations, which may be a possible regulator of leptin concentration.[49]

Insulin resistance is a prominent feature in patients with type 2 diabetes. Acarbose increases insulin sensitivity significantly in obese patients with impaired glucose tolerance.[50] Chiasson et al. reported a significantly improved insulin sensitivity (p < 0.004), measured as steady-state plasma glucose using the insulin suppression test.[51] Laube et al.[52] calculated insulin sensitivity, using the hyperglycaemic clamp test and a minimal model intravenous glucose tolerance test, in a double-blind, placebo-controlled study of acarbose in obese subjects with impaired glucose tolerance. Subjects were treated with acarbose 100mg or placebo three times daily for 3 months. Acarbose treatment caused a decrease in proinsulin/insulin ratio and a significant (p < 0.05) increase in insulin sensitivity.

Calle-Pascual et al.[53] and Meneilly et al.[34] observed an increase in insulin sensitivity of up to 30% in hyperglycaemic clamp studies in moderately obese elderly patients with type 2 diabetes receiving acarbose 50 to 100mg three times daily for 12 months. In contrast, in younger patients with poorly controlled overt type 2 diabetes, several studies have failed to show any effects of acarbose on insulin-mediated glucose disposal.[42,54,55,56] Probable reasons for this inconsistency are, firstly, the different quality of metabolic control in the two populations, and secondly, a lesser impairment of ß-cell function and lower postprandial hyperinsulinaemia in patients with less advanced disease. This also explains why acarbose is more effective in insulin-resistant, newly diagnosed patients with type 2 diabetes (figure 1).

Mechanisms of acarbose action. HbA1c = glycated haemoglobin; GLP-1 = glucagon-like peptide 1.

The mechanism by which acarbose increases insulin sensitivity is probably based on lowering fasting and postprandial hyperglycaemia and decreasing glucose toxicity.[57] In addition, a decrease in post-challenge hyperinsulinaemia is considered by some authorities to contribute to insulin sensitivity.[13] Other investigators have reported improvements in insulin sensitivity following a rise of the incretin hormone, GLP-1, and the 'priming' effect it induces (see section 1.7).[57,58]

-Glucosidase Inhibition and Intestinal Hormones

Considerable interest has recently focused on the incretin hormones. Acarbose inhibits the post-prandial release of gastric inhibitory polypeptide (GIP) in the duodenum and jejunum and increases the response of GLP-1 in the distal intestine, ileum and colon during the late postprandial period (60 to 240 min).[59,60] GLP-1 primes ß-cells and makes them more sensitive to glucose, thus increasing their secretion of insulin in response to glucose load and improving insulin sensitivity. In addition, GLP-1 delays gastric emptying and stimulates satiety. The increase in GLP-1 following acarbose treatment is therefore a reliable marker of delayed and more distal intestinal absorption of carbohydrate, and a modulator of decreased postprandial hyperglycaemia.

Treatment with acarbose is associated with several changes in lipid profile. Serum triglycerides, very low-density lipoprotein (VLDL) concentration and free fatty acids are frequently elevated in obese patients with insulin-resistant type 2 diabetes. Several studies have documented a dose-dependent reduction in blood lipids with acarbose in this patient population.[40,61,62]

Lowering of total serumtriglycerides is primarily mediated via a reduction in the biosynthesis of VLDL[61] and is secondary to acarbose-induced attenuation of postprandial hyperinsulinaemia.[63] Mean triglycerides decreased significantly (from 5.8 mmol/L to 3.6 mmol/L) when acarbose 50mg twice daily was given as an adjunct to dietary therapy in 30 nondiabetic patients with hypertriglyceridaemia for a total period of 16 weeks.[64] The same beneficial effect was seen in 18 non-diabetic patients with familial hypertriglyceridaemia (FH); mean serum triglycerides dropped significantly (p < 0.05) from 5.8 ± 4.1 to 3.6 1.2 mmol/L after 2 months' treatment with acarbose 50mg twice daily.[65]

The response of fasting triglyceride levels to acarbose is related to dietary fat intake and an overall improvement of metabolic control.[42] Maruhama et al.[66] reported a significant (p < 0.05) mean decrease in fasting serum triglycerides (from 1.92 ± 0.31 mmol/L to 145 ± 0.21 mmol/L) in obese hyperinsulinaemic patients following acarbose 100mg three times daily for 1 month.

Carbohydrate-induced postprandial triglyceride overproduction is reduced for several hours by acarbose, through a slowing of the impact of glucose on liver metabolism.[30,67] Similar results were reported by Kado et al.,[68] who demonstrated a significant (p < 0.01) reduction of the post-prandial rise of serumtriglycerides and lipoprotein remnants in the postprandial phase in 20 normal weight patients with type 2 diabetes, following a 300 kcal test meal (21.1% protein, 22.5% fat, 49.6% carbohydrate) and a single dose of acarbose 100mg.

Since acarbose does not interfere with intestinal lipid absorption, the most likely mechanism for its hypotriglyceridaemic action is a slower hepatic uptake of precursor molecules for de novo lipogenesis. Dietary carbohydrates are key precursors of lipogenesis and insulin plays a central role in postprandial lipid metabolism. Thus, acarbose may also contribute to triglyceride inhibition by interference with endogenous triglyceride synthesis. Suppression of intestinal lipogenesis by acarbose has also been suggested as a plausible explanation.[68]

Inconsistent effects of acarbose on serum cholesterol have been reported. Total cholesterol concentrations were not significantly altered in studies reported by Homma et al.[69] and Nestel et al.,[61] whereas other studies have documented a significant reduction.[70,71,72] An increase in the low-density lipoprotein/high-density lipoprotein (LDL/HDL) cholesterol ratio of 26.8% was evident following treatment of 96 patients with type 2 diabetes with acarbose 100mg three times daily for 24 weeks in the Essen-II Study.[73] Plasma levels of apolipoprotein A-I and A-II decreased significantly during acarbose treatment, whereas plasma apolipoprotein B remained unchanged.[74]

In hyperinsulinaemic, overweight patients with impaired glucose tolerance, acarbose 300mg daily reduced LDL-cholesterol significantly (p < 0.05) from 4.40 ± 0.30 mmol/L to 3.40 ± 0.27 mmol/L after 4 weeks. HDL-cholesterol remained unchanged.[70] There was a marked increase of intestinal anaerobic bacteria (bifidobacter and acidophilus), probably as a result of undigested carbohydrates in the lower part of the bowel.

Significantly elevated levels of ursocholic acids in the stool appear to be the additive consequence of a decreased rate of absorption and increased intestinal motility due to the changes of intestinal bacteria. Lack of biliary acids, however, induce accelerated LDL-cholesterol uptake in the liver and decrease serum cholesterol. The marked correlation between the amount of faecal bifido bacteria, faecal biliary acid level and LDL-cholesterol points to an underlying mechanism for the decreased serum cholesterol following acarbose.[70]

A reduction in serum cholesterol has also been attributed to a lowered VLDL concentration and a diminished conversion rate of VLDL to LDL following acarbose treatment. Acarbose affects the activity of HMG-CoA reductase, essential for the biosynthesis of cholesterol,[75] by stimulating significantly the excretion rate of cholic acids in the stool.[70]

In overweight patients with type 2 diabetes and mild hypertension, acarbose lowers systolic blood pressure (5.2 ± 2.4mm Hg) significantly (p = 0.0001) after 24 weeks and decreases heart rate slightly compared with glibenclamide [76] These effects can be explained by the lower insulin levels following acarbose treatment, which, in turn, reduce sympathetic nervous system activity and insulin- induced vasodepressor action.

Patients with diabetes mellitus are prone to thrombosis, particularly since postprandial hyperglycaemia activates haemostasis. In 17 patients with type 2 diabetes maintained on diet therapy alone who consumed a standard meal (372 kcal, 49% carbohydrate, 40% fat, 11% protein), a single dose of acarbose 100mg significantly attenuated the postprandial rise of prothrombin fragments 1 and 2 (2.0 vs 2.7 mmol/L at 2 hours) and D-dimer (2.85 vs 3.50 g/L at 1 hour) compared with placebo. [77] These components are sensitive markers of ongoing coagulation activation and fibrinolysis. Prothrombin fragments 1 and 2 have been shown to be strong predictors of thrombotic coronary stent occlusion, and D-dimer, a primary degradation product of cross-linked fibrin, is a predictor of myocardial infarction.[78] Thus, acarbose may be useful in reducing meal-induced activation of haemostasis in the procoagulative state of diabetes mellitus.


Comments on Medscape are moderated and should be professional in tone and on topic. You must declare any conflicts of interest related to your comments and responses. Please see our Commenting Guide for further information. We reserve the right to remove posts at our sole discretion.