Review

CLINICAL EFFICACY OF ACARBOSE IN DIABETES MELLITUS : A CRITICAL REVIEW OFCONTROLLED TRIALS

A.J. SCHEEN

Summary: Acarbose, an alpha-glucosidase inhibitor, is a new antihyperglycaemic agent which hasbeen proposed as add-on therapy in Type 2 diabetic patients not well-controlled withdiet alone, sulphonylurea, metformin or insulin, and in Type 1 diabetic patients withlarge meal-related plasma glucose excursions. Numerous controlled studies investigatingthe clinical effects of acarbose in Type 2 diabetes versus either placebo or,more rarely, versus a reference drug (sulphonylurea or metformin) have beenpublished during the last 10 years. All placebo-controlled studies have demonstratedthe superiority of acarbose, at a dose of 150-600 mg/day, in decreasing fasting andpostprandial glucose levels as well as HbA1c concentrations (mean decrease of0.7 %), whether acarbose was given as first-line therapy in diet-treated diabeticpatients or in combination in individuals already receiving a sulphonylurea, metformin orinsulin. Only a few controlled studies have compared the effects of acarbose with those ofeither sulphonylurea or metformin, yielding controversial results. In Type 1 diabeticpatients, a small reduction of HbA1c levels was also reported after addition ofacarbose to insulin therapy, which in some cases allowed a slight reduction of dailyinsulin needs. All these favourable biological effects occurred without exposing thepatient to hypoglycaemia or weight gain. A few studies have also reported favourableeffects on postprandial lipid profile and some other vascular risk factors. However, it isnot clear whether the extra cost of acarbose, when compared to that of older oralantidiabetic agents, is justified since no study has yet demonstrated its potentialbenefit on the complications and long-term prognosis of diabetic patients.

Key-words: alpha-glucosidase inhibitor, Type 1 diabetes, Type 2 diabetes, glucose control,postprandial, hyperinsulinism, dyslipidaemia.

Efficacité clinique de l'acarbose dans lediabète sucré : une revue critique des essais contrôlés.

Résumé: L'acarbose, un inhibiteur des alpha-glucosidases digestives, est un nouveau médicamentantihyperglycémiant proposé comme thérapie complémentaire chez les patientsdiabétiques de type 2 insuffisamment contrôlés par régime seul, sulfamide,metformine ou insuline, et chez les patients diabétiques de type 1 présentant delarges fluctuations glycémiques liées aux repas. De nombreuses études contrôléesanalysant les effets cliniques de l'acarbose dans le diabète de type 2 versussoit un placebo, soit une molécule de référence (sulfamide ou metformine) ont étépubliées au cours des 10 dernières années. Toutes les études contrôlées versusplacebo démontrent la supériorité de l'acarbose pour diminuer la glycémie à jeun etpost-prandiale ainsi que le taux d'HbA1c (diminution moyenne de 0,7 %),que l'acarbose soit donné en première intention après échec du régime ou encombinaison chez des patients déjà traités par un sulfamide, plus rarement lametformine ou l'insuline. Assez peu d'études contrôlées ont comparé les effets del'acarbose avec ceux d'une sulfonylurée ou de la metformine et elles ont donné desrésultats contradictoires. Chez les patients diabétiques de type 1, une légèreréduction du taux d'HbA1c a aussi été rapportée après l'additiond'acarbose à l'insulinothérapie, avec dans certains cas une discrète diminution desbesoins insuliniques journaliers. Tous ces effets biologiques favorables surviennent sansexposer le patient à un risque accru d'hypoglycémie ni de gain de poids. Quelques raresétudes ont également rapporté des effets favorables en phase post-prandiale sur leprofil lipidique et d'autres facteurs de risque vasculaire. Il reste cependant à prouverque le surcoût occasionné par l'introduction de cette nouvelle molécule dans letraitement du diabète est bien justifié puisqu'aucune étude n'a encore démontré lebénéfice potentiel de ce médicament en ce qui concerne les complications et lepronostic à long terme des patients diabétiques.

Mots-clés: Inhibiteur des alpha-glucosidases, diabète de type 1, diabète de type 2, post-prandial,hyperinsulinémie, dyslipidémie, contrôle glycémique.

Acarbose, an alpha-glucosidase inhibitor, has been proposed as an alternative drug to,or in combination with, classical antidiabetic agents such as sulphonylureas, metforminand insulin in the treatment of diabetes mellitus [1-3]. Considering its original mode ofaction [4] and the potential beneficial effect on postprandial hyperglycaemia, it may behelpful in Type 2 as well as Type 1 diabetic patients, regardless of the form ofconcomitant treatment [5-10]. However, the precise role of acarbose in the overalltherapeutic strategy of diabetes mellitus remains controversial, and the problem of theextra cost due to the large prescription of this new and more expensive drug has beenquestioned [11-14]. A better understanding of what could be expected, from a biologicaland clinical point of view, by the addition of acarbose to the treatment of variouscategories of diabetic patients should help the clinician to define the role of this drugmore precisely in the treatment of diabetes mellitus in daily clinical practice [15-19].However, in the case of acarbose, as classically reported with other drugs, the publishedresults of open studies [20] or post-marketing surveys [21] appear to be more favourablethan those obtained in carefully controlled trials [19].

The present review aims at summarising the therapeutic effects of acarbose according tothe data of numerous controlled clinical trials in the literature, especially for Type2 diabetes. The efficacy of acarbose in Type 1 diabetes will also bebriefly considered, as well as the possible additional effects of the drug which maycontribute to improving the long-term prognosis for diabetic patients.

GLYCAEMIC CONTROL IN TYPE 2 DIABETES

Numerous non-controlled studies performed in Type 2 diabetic patients havedetermined that acarbose improves glycaemic control, not only in the postprandial statebut also in fasting conditions, as confirmed by significant reductions in HbA1clevels [5-10]. In a large post-marketing survey performed in Germany on4,071 Type 2 diabetic patients, the addition of acarbose to either dietalone, sulphonylurea, insulin or insulin + sulphonylurea resulted in asignificant decrease of HbA1c levels, which averaged 1 % in individualswith acceptable glycaemic control and 2.3 % in subjects with very poor initialcontrol [21]. These results were confirmed by an interim report of the PROTECT trialperformed in the United States in which the decrease in HbA1c level averaged0.8 % 28 weeks after addition of acarbose (progressive titration from3 x 25 mg/day up to 3 x 100 mg/day) in2,139 diet-treated (23 % of the individuals) or sulphonylurea-treated(77 %) Type 2 diabetic patients [20]. The efficacy of acarbose wasindependent of concomitant antidiabetic drug treatment, age, body weight and race, but, asin the German survey [21], clearly depended on the initial quality of blood glucosecontrol : the higher the HbA1c level, the greater the reduction observedwith acarbose [20]. Furthermore, the efficacy of acarbose may be strongly influenced bythe content of carbohydrate in the diet [22], as recently indicated once again by aJapanese open study [23]. Indeed, no significant reduction in HbA1c levelscould be detected after 6 months of acarbose 300 mg/day in sulphonylurea- orinsulin-treated Type 2 diabetic patients with low carbohydrate intake(< 50 % of energy intake as carbohydrates), which contrasted with a clear-cutreduction averaging 1.5-2.0 % in the subgroup with high (> 50 %)carbohydrate intake.

We will focus our analysis here only on carefully controlled studies, either versusplacebo or versus a reference drug (sulphonylurea or metformin), published ininternational journals. All these studies have evaluated acarbose as monotherapy indiet-treated diabetic patients (Table I)(Voir Tableau)or in combination with another classical antidiabetic treatment in patients alreadyreceiving a sulphonylurea, metformin or insulin (Table II)(VoirTableau) . It is noteworthy that some of these studies were performed on quite smallnumbers of patients followed during a rather short period of time. Furthermore, availableresults have almost never been analysed according to the "intention-to-treat"statistical procedure, despite the fact that this method has been recommended forevaluation of clinical trials with many drop-out patients. This may be important in thecase of acarbose for which patient compliance may be hindered because of commongastrointestinal side-effects. One rare exception was the recent but preliminary report ofthe United Kingdom Prospective Study (UKPDS) [24]. In this placebo-controlled double-blindtrial performed on 1,946 Type 2 diabetic patients treated with diet alone(14 %), sulphonylureas (26 %), metformin (6 %), a sulphonylurea-metformincombination (16 %) or insulin (28 %), administration of acarbose (up to3 x 100 mg/day) was associated with a significant reduction of HbA1clevels of 0.6 % in the "actual therapy" analysis, whereas this decrease wasreduced by half (only 0.3 %, although highly significant) with the"intention-to-treat" analysis. Thus, these results were much less impressivethan those of the two previously mentioned open studies [20, 21], possibly because ofbetter initial metabolic control in the UKPDS study (initial HbA1c levelsaveraging 7.5 %) [24]. Interestingly, the favourable effects of acarbose on glycaemiccontrol were again similar, regardless of the concomitant antidiabetic drug treatment.

Acarbose in diet-treated patients

Controlled studies versus placebo - Several controlledstudies have demonstrated the favourable effects of acarbose as first-line therapy afterdiet failure in Type 2 diabetic patients (review in [25]). The most importantstudies, with a follow-up of at least 16 weeks, are summarised in Table I[26-37]. All were performed according to a parallel design, and most of them used a doseof 3 x 100 mg acarbose per day, sometimes after a initial transient periodof 3 x 50 mg/day. When compared to placebo, the administration of acarboseinduced a significant reduction in fasting plasma glucose levels (0.6 to2.1 mmol/l), postprandial plasma glucose concentrations (1.2 to3.6 mmol/l), and HbA1c levels (0.2 to 1.3 %). The mean HbA1cdecrease calculated on a cohort of 410 patients [26-35] averaged 0.67 % (Table I).

Controlled studies versus sulphonylurea - During the16 weeks following withdrawal of sulphonylurea therapy in patients withType 2 diabetes mellitus, a marked deterioration of glycaemic control wasobserved with placebo, and this metabolic worsening was practically unreduced in9 patients receiving acarbose (150-300 mg/day) (increase of HbA1clevels from 10.5 to 12.4 %, p < 0.001) [26]. This preliminary studysuggested that acarbose was not an effective substitute for sulphonylureas in non-obesepatients with Type 2 diabetes uncontrolled by diet alone. Only a few controlledstudies have compared the efficacy of acarbose with that of a sulphonylurea compound indiet-failure Type 2 diabetic patients. Controversial results have been reported,with either quite similar antihyperglycaemic effects with acarbose (300 mg/day) andglibenclamide (3.5 to 10.5 mg/day) [31, 36, 37] (all studies from the sameGerman group), but poorer results with acarbose (despite a high dosage of3 x 200 mg/day) as compared to tolbutamide [38] (Table I).

Controlled studies versus metformin - To our knowledge,and quite surprisingly, only one study (reported recently in abstract form [35]), hascompared the efficacy of acarbose 3 x 100 mg/day with that of metformin2 x 850 mg/day in diet-failure Type 2 diabetic patients. Resultsafter 24 weeks showed similar glycaemic control with the two drugs (Table I).Acarbose exerted more favourable effects on the lipid profile (see below) but caused moregastrointestinal side-effects than metformin.

Acarbose in sulphonylurea-treated diabetic patients

Most trials performed in sulphonylurea-treated Type 2 diabetic patients haveattempted to improve metabolic control in poorly controlled individuals [30, 38-47].However, to our knowledge, no study has clearly demonstrated that the addition of acarboseleads to a significant delay in progression to insulin therapy after secondary failure toclassical oral therapy. Furthermore, no placebo-controlled trials have investigated theeffects of adding a sulphonylurea agent in Type 2 diabetic patients receivingacarbose as first-line therapy.

Controlled studies versus placebo - Several studies [30,39-43] carried out according to a parallel or a cross-over protocol design have shown thatacarbose in combination with a sulphonylurea compound can provide a significant reductionof HbA1c levels (0.3 to 1.2 %), fasting glucose concentration(0.1 to 2.24 mmol/l) and post-meal glycaemia (0 to 2.9 mmol/l) whencompared with placebo. The mean HbA1c decrease calculated on a cohort of259 patients [30, 38, 39, 41-43] averaged 0.73 % (Table II).

Controlled studies versus metformin - Some older studiescompared the effects of acarbose with those of phenformin or buformin inType 2 diabetic patients not well-controlled with sulphonylurea treatment(review in [48]). As these two biguanide agents have been withdrawn from the marketbecause of excessive risk of lactic acidosis, only comparative trials with metformin willbe considered in the present review. Although acarbose allowed better control ofpostprandial hyperglycaemia as compared to metformin (the latter, however, was used at arather small dosage in some studies), the differences between the two drugs were notconsistent in the fasting state, and controversial results have been reported regardingthe comparative effects on HbA1c levels (Table II).

Acarbose in metformin-treated diabetic patients

To our knowledge, only one long-term placebo-controlled study has investigated theeffect of adding acarbose to a treatment with metformin in Type 2 diabeticpatients [30]. This multicentre Canadian trial showed a more marked HbA1creduction of 0.8 % with acarbose than with placebo : the difference was quitemodest (1.03 mmol/l), although statistically significant, for postprandial glycaemia,but slightly greater (1.5 mmol/l), though not significant, for fasting plasma glucoselevels (Table II).

Another short-term trial, which was recently reported in abstract form [48],investigated the effects of adding metformin as second-line drug treatment in Type2 diabetic patients already treated with acarbose but poorly controlled withmonotherapy. This combination allowed a significant improvement in the parameters of bloodglucose control : after 12 weeks, when compared to placebo, a significantreduction of HbA1c level (1.6 %), fasting glycaemia (0.8 mmol/l) andpostprandial glycaemia (2.1 mmol/l) was observed [49]. Thus, despite the fact thatacarbose may reduce the bioavailability of metformin to some extent [50], a clear-cutpotentiation of the antihyperglycaemic effect could be obtained inType 2 diabetic patients by combining these two antidiabetic oral agents.

Acarbose in patients with a sulphonylurea-biguanide combination

In a placebo-controlled study, no significant improvement in HbA1c levelswas detected in 15 Type 2 diabetic patients receiving 150 mg/day ofacarbose for 4 months in addition to maximal oral treatment with sulphonylurea andbiguanide [51]. The authors suggested that these disappointing results may have been dueto the relatively low dosage of acarbose used, the importance of carbohydrate restriction(only 40 % of the energy intake) and/or the selection of patients since most of themmight have been classified as insulin-requiring diabetic patients. An open study performedin Thailand [52] reported that acarbose 300 mg/day significantly reduced HbA1clevels (p < 0.001) but allowed values below 8 % to be reached in only17 % of the 36 Type 2 diabetic patients poorly controlled withsulphonylurea alone or in combination with metformin. As body weight also decreasedsignificantly, it is difficult to decide whether the positive results were attributable toacarbose or diet reinforcement. An open pilot study recently showed a positive effectthrough addition of acarbose to a sulphonylurea-metformin combined therapy [53]. In arather small number of 11 patients with very poor control on classical combined drugtreatment, the addition of acarbose at a dose of 3 x 50-100 mg/day alloweda moderate but significant reduction in HbA1c levels from 13.6 to12.2 %. Thus, the potential interest of such triple therapy should be confirmed in acontrolled study on a larger number of Type 2 diabetic subjects, if possiblewith much better initial glycaemic control. Indeed, in the preliminary results of theUKPDS placebo-controlled study [24], the improvement in fasting plasma glucose and HbA1clevels observed after addition of acarbose was significant and similar in patientsreceiving a combination therapy to that in patients with monotherapy (sulphonylurea ormetformin).

Acarbose in insulin-treated Type 2 diabetic patients

In an open short-term (one week) pilot study performed in Japan, acarbose(300 mg/day) improved daily glucose profiles and possibly reduced insulin requirementin 9 Type 2 diabetic patients receiving insulin therapy [54]. In arandomised placebo-controlled parallel study, a moderate but significant reduction in HbA1clevels (- 1.5 % in the acarbose-treated group versus - 0.9 %in the placebo-treated group) was reported in Type 2 diabetic patients receivingat least 20 IU insulin/day [55]. Besides this study, only two placebo-controlledlong-term studies have compared the effects of acarbose in combination with insulintherapy in Type 2 diabetic patients [30, 56]. Both reported concordant resultswith a modest, but significant, reduction of HbA1c levels (0.4 %) despitea decrease in daily insulin dosage of about 15 % after 24 weeks [56] or of30 % after 56 weeks [30]. Finally, a recent placebo-controlled study publishedonly in abstract form showed in 48 Type 2 diabetic patients that additionof acarbose to insulin therapy after secondary failure to sulphonylurea led to asignificant increase in the number of good responders to insulin after 24 weeks (andperhaps even at a lower insulin dose) [57].

GLYCAEMIC CONTROL IN TYPE 1 DIABETES

Very early in its clinical development, acarbose was evaluated inType 1 diabetic patients [58]. Most initial studies were interested indemonstrating the acute or short-term (up to several weeks) effects of acarbose in thispopulation (review in [59, 60]). Some studies showed the insulin-sparing effect ofacarbose when an artificial pancreas was used (Biostator^®) (review in [12]).Most observations dealing with the use of acarbose in Type 1 diabetic patientshave been reported in the proceedings of three symposia on acarbose [5-7], and rather fewstudies have been published in peer-reviewed journals (Table III)(Voir Tableau) [61-65]. Almost all clinical studies have shown that theaddition of acarbose to insulin dampened postprandial glucose elevations, reduced the riskof delayed hypoglycaemia, especially at night [66], "smoothed" the daily plasmaglucose profile, with a reduction of early postprandial peak levels and delayed nadirvalues (resulting in a diminution of the "M" and "MAGE" indices ofinstability), lowered mean daily glucose and HbA1c levels, and somewhatdecreased daily insulin needs (review in [59, 60]). Unfortunately, few studies havecovered a long time-period (follow-up limited to 4-12 weeks, except for a singlerecent study of 24 weeks) or been conducted according to a double-blind randomisedprotocol (Table III) [61-65]. To date, no study has demonstrated theusefulness of such insulin-acarbose combined treatment in the long-term, especially inpreventing or retarding degenerative diabetic complications. Finally, even though acarbosehas been shown to improve indices of glycaemia instability, it has not yet proven its realefficacy in Type 1 patients with brittle diabetes.

OTHER BENEFICIAL EFFECTS OF ACARBOSE

In recent years, several studies have emphasised the deleterious role of biologicalabnormalities occurring in the postprandial state, which could increase the risk ofcardiovascular disease in both diabetic and non-diabetic individuals [67]. As acarbosespecifically acts by interfering with the absorption of nutrients following a meal,several beneficial effects may be expected in addition to a pure antihyperglycaemiceffect.

Effects on the lipid profile - Several animal studies havesuggested that high dosages of acarbose could reduce serum lipid parameters [4].Preliminary studies in healthy volunteers and initial open trials in diabetic subjectsappeared to confirm these findings and showed a significant decrease in fasting andpostprandial triglyceride levels (about 0.3 mmol/l) [52, 58, 59, 63]. However, mostcontrolled studies failed to detect any significant changes in fasting total cholesterol,HDL cholesterol and triglyceride levels in both Type 2 andType 1 diabetic patients (Table IV)(Voir Tableau). In one study, a clear-cut reduction in total cholesterol levels was observed only in asubgroup of Type 2 diabetic subjects with initial hypercholesterolaemia above6.66 mmol/l or 260 mg/dl [34, 70]. Very few studies have investigated thepostprandial state [27, 68]. One placebo-controlled trial demonstrated a mild butsignificant reduction (- 0.5 mmol/l, p < 0.05) in post-mealtriglyceride levels in Type 2 diabetic patients [27, 71]. Finally, opposite resultshave been recently reported as far as the effects of acarbose on Lp(a) and ApoB levels areconcerned in Type 2 diabetic patients [72]. Indeed, administration of acarbose forone year was accompanied by a modest but significant increase in Apo-B and no change inLp(a) concentrations, while placebo treatment was associated with no change in Apo-B but asignificant increase in Lp(a) levels. These results still need to be confirmed in a largercohort before any conclusions can be drawn about their possible clinical significance.Thus, the long-term effects of acarbose on lipid parameters appear rather small, at leastin the fasting state and with the dosages used in clinical practice, and their potentialimpact on atherogenic risk remains doubtful.

Effects on insulin sensitivity - It has been suggested thatacarbose, both by decreasing hyperglycaemia (reduction of glucotoxicity) andhyperinsulinaemia (less pronounced down-regulation of insulin receptors) during the hoursfollowing a meal, could improve insulin sensitivity. Such an effect on insulin action hasnot been observed in Type 2 diabetic patients with frank hyperglycaemia [69,73]. However, a partial correction of insulin resistance was reported at the hepatic butnot the muscular site in a 2-week open study in diabetic patients [68], inType 2 diabetic patients in whom the treatment with acarbose was accompanied bya significant weight loss [34] or, more recently, in non-diabetic subjects with impairedglucose tolerance [74].

Effects on vascular risk factors - It has been reported thatsome vascular risk factors frequently observed in Type 2 diabetic patients can beimproved by acarbose [34, 75]. In a large placebo-controlled Australian study performed on105 individuals, acarbose at a dose of 300 mg/day for 16 weeks reducedfasting glucose and triglyceride levels, lowered HbA1c and limited insulin andproinsulin responses to a standard meal in Type 2 diabetic patients withplurimetabolic syndrome, although it did not affect fasting total and HDL cholesterollevels, fibrinogen concentrations and blood pressure values [75]. Finally, a cross-overplacebo-controlled study demonstrated that acute ingestion of 100 mg acarbose led toa significant reduction in coagulation markers which are increased in the postprandialstate in Type 2 diabetic patients [76]. The clinical significance of theseeffects remains to be elucidated.

Effects on diabetic complications - From a theoretical point ofview, acarbose should be able to improve the prognosis of diabetic microangiopathy byinducing sustained decrement of plasma glucose levels (as assessed by a mean 0.7 %decrease in HbA1c levels in published placebo-controlled studies) (Tables Iand II), and that of macroangiopathy by exerting favourable effects on glucose,insulin and the postprandial lipid profile, and possibly on coagulation as well [77]. Infact, studies on diabetic animals have shown the potential of acarbose therapy inpreventing or delaying the development of diabetic microangiopathies [2]. No humanstudies, however, have shown direct or indirect evidence that acarbose can decrease theprevalence, incidence or severity of diabetic vascular complications [14, 63]. The UKPDSstudy [24] will perhaps provide such evidence in the near-future.

CONCLUSIONS

Numerous controlled studies have investigated the antihyperglycaemic effects ofacarbose in Type 2 and to a lesser extent in Type 1 diabetic patients.Almost all studies have demonstrated that acarbose is significantly more effective thanplacebo in reducing fasting and postprandial plasma glucose levels and HbA1cconcentrations. These antihyperglycaemic effects occurred without increasing (andsometimes by decreasing) the risk of hypoglycaemic episodes or of weight gain. Someadditional favourable effects on the lipid and vascular risk profile in the postprandialstate have also been reported in Type 2 diabetic patients. Theantihyperglycaemic effect of acarbose was observed in patients treated with diet alone,sulphonylurea, metformin or insulin. Fewer studies have compared the effects of acarbosewith those of other oral antidiabetic agents such as sulphonylureas or metformin.Furthermore, the results are more conflicting since almost similar, greater or lowerreducing effects of HbA1c levels have been described with acarbose whencompared to other classical oral drugs, especially glibenclamide or metformin. Finally, asdigestive intolerance may reduce compliance to prolonged treatment with acarbose, theresults of clinical trials should be best analysed on an intention-to-treat basis, whichhas not been the case in most published studies.

Although the antihyperglycaemic efficacy and safety of acarbose have been welldemonstrated, the precise role of acarbose in the treatment of diabetes mellitus remainscontroversial. The controversy results essentially from the higher cost of this drug whencompared to that of sulphonylureas or metformin and the absence to date of any proof ofits superiority regarding relevant clinical endpoints [12, 13, 18, 19]. Hopefully, theresults of the UKPDS, in which a large group of patients is receiving acarbose in additionto various classical antidiabetic therapies (diet, sulphonylureas, metformin or insulin)according to a placebo-controlled protocol [24], will help the clinician to determine thereal long-term advantages of acarbose, not only on surrogate end-points but also onprimary endpoints such as microangiopathy, macroangiopathy and mortality.

Note added in proofs - Since the acceptance of this manuscript, the results of alarge double-blind, placebo-controlled, multinational, five-arm study have been reported[79]. The clinical trial included 495 type 2 diabetic patients from 7 Europeancountries who were insufficiently controlled with diet alone. Acarbose was given asfirst-line drug at 4 different dosages (25, 50, 100 and 200 mg t.i.d.) versusplacebo for 24 weeks. No significant differences were observed between the fivegroups at baseline. At the end of the study (n = 420) the mean 2 hposprandial areas under the curve values for blood glucose were 22.6, 21.2, 19.6, 20.3 and18.5 mmol/l in the patients given placebo or acarbose 25, 50, 100 and 200 mgt.i.d., respectively. The corresponding HbA1c levels for the placebo andacarbose groups were 7.8 %, 7.37 %, 6.98 % and 6.79 %. Less than3 % of patients stopped tablet intake due to adverse events. In conclusion, accordingto this large dose-reponse study performed in diet-failure patients with type 2diabetes 1) small doses of acarbose (25 mg t.i.d.) had a therapeutic effect;2) a plateau phenomenon was observed at a dosage of 50-100 mg t.i.d.; 3) themost marked antihyperglycaemic effect was detected with the highest dosage of 200 mgacarbose t.i.d.; and 4) acarbose was well tolerated in the dosage range of 25 mgt.i.d. to 200 mg t.i.d., without clear relationship between the dosage and thefrequency of gastrointestinal side effects.

TABLE I.    Antihyperglycaemic activityof acarbose in diet-treated Type 2 diabetic patients : comparativerandomised parallel studies of at least 16 weeks duration controlled versus placebo,sulphonylurea or metformin.

^a Versus glibenclamide 1 to3 x 3.5 mg/day. ^b Versus glibenclamide3.5-7.0 mg/day (mean : 4.3 mg/day). ^c Versus tolbutamide250-1,000 mg 3 x /day. ^d Versus metformin2 x 850 mg/day.

TABLE II.    Antihyperglycaemic activityof acarbose in combination with classical antidiabetic drug treatment in Type2 diabetic patients : data from the literature.

^x 48 % treated with glibenclamide (mean dose of 8 mg/day). ^y 76 %of the patients treated with sulphonylureas, 24 % treated by diet alone.* diet + maximal oral treatment (sulphonylurea + biguanide).** 14 % of patients treated by diet alone, 58 % by oral agents(sulphonylureas and/or metformin), 28 % by insulin (see details in the text). ^a Versusmetformin 2 x 850 mg/day. ^b Versus metformin500 mg/day. ^c Versus metformin 850 mg/day. ^d Versusmetformin 1,500 mg/day.

TABLE III.    Antihyperglycaemic activityof acarbose versus placebo in combination with insulin therapy in Type 1 diabeticpatients : data from the literature.

TABLE IV.    Effects of acarbose versusplacebo on fasting lipid profile in Type 2 and Type 1 diabetic patients.

* Post-hoc analysis in a subgroup with initial totalcholesterol > 6.66 mmol/l (260 mg/dl).** p < 0.05 (all other differences were not significant). NS : nosignificant changes (absolute values not available in the original publication).

A.J. Scheen, Division of Diabetes, Nutrition and metabolic Disorders,Department of Medicine, CHU Sart Tilman (B35), B-4000 Liège Belgium. Tél :32-4-3667237. Fax : 32-4-3667068. Email : Andre.Scheen & chu.ulg.ac.beReceived : February 21, 1998.

REFERENCE(S)

1 Clissold SP, Edwards C. Acarbose. A preliminary review of its pharmacodynamic andpharmacokinetic properties, and therapeutic potential. Drugs, 1988, 35,214-243.

2 Balfour JA, McTavish D. Acarbose - an update of its pharmacology and therapeuticuse in diabetes mellitus. Drugs, 1993, 46, 1025-1054.

3 Scheen AJ. Drug treatment of non-insulin-dependent diabetes mellitus in the 1990s :achievements and future developments. Drugs, 1997, 54, 355-368.

4 Brendel E, Wingender W. Clinical pharmacology of glucosidase inhibitors. In :Kuhlmann J, Puls W, eds. "Oral antidiabetics. Handbook of ExperimentalPharmacology". Berlin : Springer Verlag, 1995, 611-632.

5 Creutzfeldt W, ed. Proceedings of the First International Symposium on Acarbose.Acarbose. Effects on carbohydrate and fat metabolism. Amsterdam : Excerpta Medica,1982.

6 Creutzfeldt W, ed. Proceedings of the Second International Symposium on Acarbose.Acarbose for the treatment of diabetes mellitus. Berlin : Springer Verlag, 1988.

7 Lefèbvre PJ, Standl E, eds. New aspects in diabetes. Treatment strategies withalpha-glucosidase inhibitors. Berlin, New York : de Gruyter, 1992.

8 Chiasson J-L, Gerich JE (eds). alpha-glucosidase inhibitors : a new therapeuticoption in diabetes mellitus. Drugs, 1992, 44 (Suppl 3), 1-65.

9 Dimitradis G, Raptis S. alpha-glucosidase inhibition in the treatment of diabetesmellitus. Horm Metabol Res, 1992, 26 (Suppl), 42-49.

10 Nieuwenhuijzen Kruseman AC, Sels JPJE, Wolffenbuttel BHR, editors. alpha-glucosidaseinhibitors in diabetes. Eur J Clin Invest, 1994, 24 (Suppl 3), 1-54.

11 Lerman I, Gomez Pérez FJ, Aguilar-Salinas CA, Rull Rodrigo JA. Acarbose : insearch for its real indications in current medical practice. Diabetes Care, 1996, 19,94-95.

12 Berger M, Köbberling J, Windeler J. Appraisal of effectiveness and potentialtherapeutic benefit of acarbose : a non-consensus conference. Diabetologia,1996, 39, 873-874.

13 Riddle MC. Tactics for type II diabetes. Endocrinol Metab Clin North Am,1997, 26, 659-677.

14 Scheen AJ, Lefèbvre PJ. Oral antidiabetic drugs : a guide to selection. Drugs,1998, 55, 225-236.

15 Santeusanio F, Compagnucci P. A risk-benefit appraisal of acarbose in the managementof non-insulin-dependent diabetes mellitus. Drug Safety, 1994, 11, 432-444.

16 Taylor RH, Bardoplh EM. Clinical evaluation of glucosidase inhibitors. In :Kuhlmann J, Puls W, eds. "Oral antidiabetics. Handbook of ExperimentalPharmacology". Berlin : Springer Verlag, 1996, 633-649.

17 Salvatore T, Giugliano D. Pharmacokinetic-pharmacodynamic relationships of acarbose.Clin Pharmacokinet, 1996, 30, 94-106.

18 Meng-Hee Tan. alpha-glucosidase inhibitors in the treatment of diabetes. CurrOpin Endocrinol Diab, 1997, 4, 48-55.

19 Scheen AJ. Place de l'acarbose dans le traitement du diabète sucré. Diab Metab,1998, 24, pages à indiquer (même nÝ).

20 Baron A, Neumann C, on behalf of the PROTECT Study Group. PROTECT interimresults : a large multicenter study of patients with type II diabetes. Clin Ther,1997, 19, 282-295.

21 Spengler M, Cagatay M. The use of acarbose in the primary-care setting :evaluation of efficacy and tolerability of acarbose by postmarketing surveillance study. ClinInvest Med, 1995, 18, 325-331.

22 Toeller M. Dietary treatment and alpha-glucosidase inhibitors in NIDDM. DiabetesNutr Metab, 1990, 3 (Suppl), 43-49.

23 Hara T, Nakamura J, Koh N, Sakakibara F, Takeuchi N, Hotta N. An importance ofcarbohydrate ingestion for the expression of the effect of alpha-glucosidase inhibitor inNIDDM. Diabetes Care, 1996, 19, 642-647.

24 Holman RR, Cull CA, Turner RC. Glycaemic improvement over one year in a double-blindtrial of acarbose in 1,946 NIDDM patients (Abstract). Diabetologia, 1996, 39(Suppl 1), A44.

25 Fölsch UR. Clinical experience with acarbose as first line therapy in NIDDM. ClinInvest Med, 1995, 18, 312-317.

26 Buchanan DR, Collier A, Rodrigues E, Millar AM, Gray RS, Clarke BF. Effectiveness ofacarbose, an alpha-glucosidase inhibitor, in uncontrolled non-obese non-insulin-dependentdiabetes. Eur J Clin Pharmacol, 1988, 34, 51-53.

27 Hanefeld M, Fischer S, Schulze J et al. Therapeutic potentials of acarbose asfirst-line drug in NIDDM insufficiently treated with diet alone. Diabetes Care,1991, 14, 732-737.

28 Hotta N, Kakuta H, Sano T et al. Long-term effect of acarbose on glycaemiccontrol in non-insulin-dependent diabetes mellitus : a placebo-controlleddouble-blind study. Diabetic Med, 1993, 10, 134-138.

29 Santeusanio F, Ventura MM, Contadini S et al. Efficacy and safety of twodifferent dosages of acarbose in non-insulin-dependent diabetic patients treated by dietalone. Diab Nutr Metab, 1993, 6, 147-154.

30 Chiasson JL, Josse RG, Hunt JA et al. The efficacy of acarbose in thetreatment of patients with non-insulin-dependent diabetes mellitus : a multicentercontrolled clinical trial. Ann Intern Med, 1994, 121, 928-935.

31 Hoffmann J, Spengler M. Efficacy of 24-week monotherapy with acarbose,glibenclamide, or placebo in NIDDM patients : the Essen study. Diabetes Care,1994, 17, 561-566.

32 Coniff RF, Shapiro JA, Seaton TB. Long-term efficacy and safety of acarbose in thetreatment of obese subjects with non-insulin-dependent diabetes mellitus. Arch InternMed, 1994, 154, 2442-2448.

33 Coniff RF, Shapiro JA, Robbins D et al. Reduction of glycosylated hemoglobinand postprandial hyperglycemia by acarbose in patients with NIDDM : aplacebo-controlled dose-comparison study. Diabetes Care, 1995, 18, 817-824.

34 Calle-Pascual A, Garcia-Honduvilla J, Martin-Alvarez PJ et al. Influence of16-week monotherapy with acarbose on cardiovascular risk factors in obese subjects withnon-insulin-dependent diabetes mellitus. A controlled, double blind comparison study withplacebo. Diab Metab, 1996, 22, 201-202.

35 Hoffmann J, Spengler M. Efficacy of 24-week monotherapy with acarbose, metformin orplacebo in NIDDM patients with diet failure (Abstract). Diabetologia, 1996, 39(Suppl 1), A230.

36 Fölsch UR, Spengler M, Boehme K, Sommerauer B. Efficacy of glucosidase inhibitorscompared to sulphonylureas in the treatment and metabolic control of diet treated type IIdiabetic subjects : two long-term comparative studies. Diab Nutr Metab, 1990, 3(Suppl 1), 63-68.

37 Spengler M, Hansel G, Boehme K. Efficacy of 6 months monotherapy withglucosidase inhibitor Acarbose versus sulphonylurea glibenclamide on metabolic control ofdietary treated type II diabetics (NIDDM). Horm Metabol Res, 1992, 26(Suppl), 50-51.

38 Coniff RF, Shapiro JA, Seaton TB, Bray GA. Multicenter, placebo-controlled trialcomparing acarbose (BAY g 5421) with placebo, tolbutamide, and tolbutamide-plus-acarbosein non-insulin-dependent diabetes mellitus. Am J Med, 1995, 98, 443-451.

39 Van Gaal L, Nobels F, de Leeuw I. Effects of acarbose on carbohydrate metabolism,electrolytes, minerals and vitamins in fairly well-controlled non-insulin-dependentdiabetes mellitus. Z Gastroenterol, 1991, 29, 642-644.

40 Gomez-Perez FJ, Violante R, Sanchez-Arriaga LF, Wong B, Rull JA. Eficacia ytolerancia de acarbose en pacientes diabeticos no dependientes de insulina. Rev InvestClin, 1992, 44, 77-83.

41 Wilms B. Therapie des Typ-II-Diabetes - alpha-Glucosidasehemmer(Acarbose). Akt Endokr Stoffw, 1992, 13, 51-56.

42 de Leiva A, Pinon P, Tebar J et al.. Efficacia clinica y tolerancia de laacarbosa en el tratamiento de pacientes diabeticos no dependientes de la insulina (tipoII). Med Clin (Barcelona), 1993, 100, 368-371.

43 May C. Efficacy and tolerability of stepwise increasing dosage of acarbose inpatients with non-insulin-dependent diabetes (NIDDM), treated with sulphonylureas. Diabetesund Stoffwechsel, 1995, 4, 3-8.

44 Schöffling K, Petzoldt R, Hillebrand I, Schwedes U. Comparison of metformin andacarbose treatment in non-insulin-dependent diabetics (NIDDM). A double-blind cross-overstudy. In : Genetic Environmental Interaction in Diabetes Mellitus. Melish, Hanna& Baba, eds. Amsterdam - Oxford - Princeton, Excerpta Medica,1982, 339-343.

45 Johansen K. Acarbose treatment of sulfonylurea-treated non-insulin dependentdiabetics. A double-blind cross-over comparison of an alpha-glucosidase inhibitor withmetformin. Diab Metab, 1984, 10, 219-223.

46 Calle-Pascual A, Garcia-Honduvilla J, Martin-Alvarez PJ et al. Comparisonbetween acarbose, metformin and insulin treatment in type 2 diabetic patients withsecondary failure to sulfonylurea treatment. Diab Metab, 1995, 21, 256-260.

47 Bayraktar M, Van Thiel DH, Adalar N. A comparison of acarbose versus metformin as anadjunct therapy in sulfonylurea-treated NIDDM patients. Diabetes Care, 1996, 19,252-254.

48 Pagano G, Cavallo-Perin P. Glucosidase inhibitors and biguanides in the treatment ofnon-insulin-dependent diabetic patients. Diabetes Nutr Metab, 1990, 3(Suppl 1), 69-76.

49 Hanefeld M, Bär K, Mertes G, Berlinghoff R. Metformin improves metabolic control innon-insulin-dependent diabetics with acarbose monotherapy (Abstract). Diabetologia,1997, 40 (Suppl 1), A312.

50 Scheen AJ, Ferreira Alves de Magalhaes AC, Salvatore T et al. Reduction ofthe acute bioavailability of metformin by the alpha-glucosidase inhibitor acarbose innormal man. Eur J Clin Invest, 1994, 24 (Suppl 3), 50-54.

51 Rodier M, Richard JL, Monnier L, Mirouze J. Effect of long term acarbose (Bay g5421) therapy on metabolic control of non-insulin-dependent (type II) diabetesmellitus. Diab Metab, 1988, 14, 12-14.

52 Vannasaeng S, Ploybutr S, Nitiyanant W, Peerapatdit T, Vichayanrat A. Effects ofalpha-glucosidase inhibitor (acarbose) combined with sulfonylurea or sulfonylurea andmetformin in treatment of non-insulin-dependent diabetes mellitus. J Med Assoc Thai,1995, 78, 578-585.

53 Kaye TB. Triple oral anti-diabetic therapy. Diabetes, 1997, 46(Suppl 1), 290A.

54 Hotta N, Kakuta H, Koh N et al. The efffects of acarbose on blood glucoseprofiles of type 2 diabetic patients receiving insulin therapy. Diabetic Med,1993, 10, 355-358.

55 Rybka J, Gegorova A, Zmydlena A, Jaron P. Clinical study of acarbose. Drug Invest,1990, 2, 264-267.

56 Coniff RF, Shapiro JA, Seaton TB et al. A double-blind placebo-controlledtrial evaluating the safety and efficacy of acarbose for the treatment of patients withinsulin-requiring type II diabetes. Diabetes Care, 1995, 18, 928-932.

57 Standl E, Baumgartl HJ, Fuechtenbusch M, Stemplinger J. Effect of acarbose onadditional insulin therapy of type II diabetics with late failure of sulphonylureatherapy. Diabetes, 1997, 46 (Suppl 1), 294A.

58 Gérard J, Luyckx A, Lefèbvre P. Improvement of metabolic control ininsulin-dependent diabetics treated with the alpha-glucosidase inhibitor acarbose for twomonths. Diabetologia, 1981, 21, 446-451.

59 Tattersall R. alpha-glucosidase inhibition as an adjunct to the treatment of type1 diabetes. Diabetic Med, 1993, 10, 688-693.

60 Rios MA. Acarbose and insulin therapy in type 1 diabetes mellitus. Eur JClin Invest, 1994, 24 (Suppl 3), 36-39.

61 Viviani GL, Camogliano L, Borgoglio MG, Ohnmeiss H, Adezati L. Acarbose treatment ininsulin-dependent diabetics - a double-blind crossover study. Curr Ther ResClin Exp, 1987, 42, 1-11.

62 Marena S, Tagliaferro V, Cavallero G et al. Double-blind crossover study ofacarbose in type 1 diabetic patients. Diabetic Med, 1991, 8, 674-678.

63 Austenat E. Neuere Therapiestudien mit Acarbose bei Typ-I-Diabetikern. Akt EndokrStoffw, 1991, 12, 19-24.

64 Escobar-Jimenez F, de Leiva A, Pinon P et al. Efficacia clinica y toleranciade la acarbosa en el tratamiento de pacientes diabeticos dependientes de la insulina (tipoI). Med Clin (Barcelona), 1993, 100, 488-491.

65 Hollander P, Pi-Sunyer X, Coniff RF. Acarbose in the treatment of type1 diabetes. Diabetes Care, 1997, 20, 248-253.

66 McCulloch DK, Kurtz AB, Tattersall RB. A new approach to the treatment of nocturnalhypoglycemia using alpha-glucosidase inhibition. Diabetes Care, 1983, 6,483-487.

67 Lefèbvre PJ, Scheen AJ. The postprandial state and the risk of cardiovasculardisease. Diabetic Med, 1998, in press.

68 Baron AD, Eckel RH, Schmeiser L, Kolterman OG. The effect of short term alphaglucosidase inhibition on carbohydrate and lipid metabolism in type II(non-insulin-dependent) diabetics. Metabolism, 1987, 35, 409-415.

69 Reaven GM, Lardinois CK, Greenfield MS, Schwartz HC, Vreman HJ. Effect of acarboseon carbohydrate and lipid metabolism in NIDDM patients poorly controlled by sulfonylureas.Diabetes Care, 1990, 13 (Suppl 3), 32-36.

70  Leonhardt W, Hanefeld M, Fischer S, Schulze J, Spengler M. Beneficial effectson serum lipids in non-insulin-dependent diabetics by acarbose treatment. Arzneimittelforschung,1991, 41, 735-738.

71 Leonhardt W, Hanefeld M, Fischer S, Schulze J. Efficacy of alpha-glucosidaseinhibitors on lipids in NIDDM with moderate hyperlipidaemia. Eur J Clin Invest,1994, 24 (Suppl 3), 45-49.

72 Hegele RA, Connelly PW, Palmason C, Jenkins DJA, Wolever TMS. Differential responsesof plasma lipoprotein(a) and apolipoprotein B in NIDDM subjects treated with acarbose. DiabetesCare, 1995, 18, 272-273.

73 Jenney A, Proietto J, O'Dea K, Nankervis A, Trainedes K, D'Embden H. Low doseacarbose improves glycemic control in NIDDM patients without changes in insulinsensitivity. Diabetes Care, 1993, 16, 499-502.

74 Chiasson JL, Josse RG, Leiter LA et al. The effect of acarbose on insulinsensitivity in subjects with impaired glucose tolerance. Diabetes Care, 1996, 19,1190-1193.

75 Scott R, Lintot C. Acarbose improves the metabolic environment of NIDDM patientswith plurimetabolic syndrome. Diabetes, 1997, 46 (Suppl 1), 294A.

76 Ceriello A, Taboga C, Tonutti L et al. Post-meal coagulation activation indiabetes mellitus : the effect of acarbose. Diabetologia, 1996, 39,469-473.

77 Wolffenbuttel BH, van Haeften TW. Prevention of complications innon-insulin-dependent diabetes mellitus (NIDDM). Drugs, 1995, 50, 263-288.

78 Zimmerman BR. Preventing long-term complications. Implications for combinationtherapy with acarbose. Drugs, 1992, 44 (Suppl), 54-60.

79 Fisher S, Hanefeld M, Spengler M, Boehme K, Temelkova-Kurktschiev T. European studyon dose-response relationship of acarbose as a first-line drug in non-insulin-dependentdiabetes mellitus : efficacy and safety of low and high doses. Acta Diabol,1998, 44, 34-40.