Alpha-glucosidase inhibitors

Observational studies

In a 2-year study of the tolerability and safety of acarbose in 2035 patients the incidence of adverse effects was 7.5% and of withdrawals 2.5% [ ]. Of 1907 patients, 444 (23%) reported one or more adverse events. In 143 patients the physician considered that there was a probable or possible relation between the adverse event (all gastrointestinal) and acarbose. There were 77 deaths, but none was considered to be related to acarbose; 52 stopped taking acarbose because of an adverse event and 45 were considered to be related to acarbose. Laboratory analyses were all within the reference ranges. HbA _1c fell by 1.92%.

The addition of metformin to acarbose in 49 patients produced a synergistic effect [ ].

Of 1027 patients, 283 used acarbose as the treatment of choice [ ]. In 250 cases the physician was not sure of the benefit; 124 of these patients took acarbose and 126 patients took placebo besides regular therapy. In those taking acarbose HbA _1c fell. The adverse effects were bloating, flatulence, abdominal cramps, and diarrhea; there were moderate increases in serum transaminases.

In a post-marketing surveillance study in 27 803 patients with diabetes mellitus (94% type 2), data were reported after 12 weeks. The doses of acarbose were low: 4.1% took less than 100 mg/day, 64% 100–250 mg/day, 32% 250–300 mg/day, and 0.1% more than 300 mg/day. Only 2.1% stopped therapy, mainly because of gastrointestinal adverse events. Tolerability appeared to be good and independent of age. Abnormal liver function was reported in 0.01%. The difference between these results and those of many controlled trials may in part be explained by the fact that higher doses have been used in most trials [ ].

Reduction of insulin resistance in impaired glucose tolerance, as found in a Chinese multicenter study [ ], can be another incentive to start treatment with alpha-glucosidase inhibitors. Various studies of the effects of alpha-glucosidase inhibitors on hyperglycemia and insulin resistance in type 2 diabetes and impaired glucose tolerance have been reviewed [ ].

Comparative studies

Voglibose and acarbose have been compared in 32 patients insufficiently treated by diet in an open crossover study [ ]. The metabolic results were identical. There were fewer adverse reactions in those who took voglibose. There was increased flatulence with acarbose in 96% and with voglibose in 57%; abdominal distension was reported in 17% and 10% respectively.

In a comparison of voglibose and acarbose in 21 in-patients with type 2 diabetes who took part in a randomized crossover study of acarbose 150 mg/day and voglibose 0.9 mg/day, there was marked interindividual variation in response [ ]. For both drugs efficacy was better in those with gastrointestinal adverse effects, such as abdominal distention and flatulence.

In an open study in 57 patients acarbose and gliclazide had the same effects on HbA _1c , blood glucose, and lipids, but the ratio of HDL to LDL cholesterol increased with acarbose [ ]. Acarbose caused flatulence in 30% and diarrhea in 3% and gliclazide caused at least one mild attack of hypoglycemia in 10%.

Placebo-controlled studies

The effect of adding acarbose (maximum 100 mg tds) or placebo to insulin [ ] or metformin [ ] has been investigated in 1946 patients with type 2 diabetes. The results were comparable with the results of the UK Prospective Diabetes Study [ ]. After 3 years, 39% were still using acarbose compared with 58% using placebo. The main reasons for stopping were flatulence (30% versus 12%) or diarrhea (16% versus 8%). After 3 years the HbA _1c concentration was 0.5% lower (median 8.1% versus 8.6%). Acarbose was equally effective when added to diet, sulfonylurea, metformin, or insulin.

When acarbose or placebo was given to patients with type 1 diabetes taking insulin, acarbose reduced postprandial blood glucose but there was no difference in HbA _1c ; the only adverse effects were gastrointestinal [ ].

In a double-blind, placebo-controlled study in 74 patients for 2 years, acarbose 100 mg tds, after a stepwise increase in dosage over 5 weeks, improved HbA _1c (− 1.71%) more than placebo [ ]. Two patients taking acarbose withdrew with drug-related adverse effects.

In a randomized, double-blind, placebo-controlled, crossover study 12 healthy subjects took acarbose 100 mg or voglibose 0.3 mg tds [ ]. Postprandial glucose, the rise in plasma immunoreactive insulin, and urinary immunoreactive C-peptide were higher with acarbose than voglibose. The flatus score was higher with acarbose than voglibose, but the stool score was not different and was higher than with placebo. Voglibose before the evening meal may improve nocturnal hypoglycemia during intensive insulin therapy [ ].

Acarbose reduced insulin resistance in 192 patients over 65 years of age (mean age 70) in a double-blind, placebo-controlled study [ ]. HbA _1c was significantly but modestly reduced. The most frequent adverse effect was flatulence, which caused 12 patients (9 taking acarbose and 3 taking placebo) to withdraw.

In a multicenter, double-blind, placebo-controlled study, 81 patients, in whom treatment with metformin was inadequate, received extra acarbose or placebo during 24 weeks after a 4-week run-in period to establish the optimal dose of acarbose [ ]. HbA _1c was reduced by 1.02% and fasting blood glucose by 1.13 mmol/l. Gastrointestinal adverse effects were more common in the acarbose group.

In a placebo-controlled study of 154 patients taking glibenclamide or metformin, miglitol (starting at 25 mg tds and increasing to 50 or 100 mg tds for 24 weeks) caused more meteorism, flatulence, and diarrhea [ ]. When miglitol was added to metformin, HbA _1c improved and there was weight loss. In another study in 318 patients, more of the patients taking miglitol only or miglitol plus metformin withdrew because of flatulence and diarrhea than in the other groups [ ].

The effects of miglitol 25 or 50 mg tds have been compared with placebo and various doses of glibenclamide in 411 patients aged over 60 years (mean 68 years) with mild type 2 diabetes insufficiently controlled by diet for 56 weeks [ ]. HbA _1c fell after 1 year by 0.92% (glibenclamide), 0.49% (miglitol 25 mg), and 0.40% (miglitol 50 mg). Gastrointestinal events were most common in patients taking miglitol. Most (88%) of the hypoglycemic events were minor and occurred in the patients taking glibenclamide. Mean body weight increased continuously and significantly in those taking glibenclamide. In the other groups weight fell by more than 1 kg. The rate of withdrawal was the same in all groups; withdrawal was mostly occasioned by cardiovascular effects in those taking glibenclamide and by hyperglycemia, flatulence, or diarrhea in those taking miglitol.

When patients with inadequately controlled type 2 diabetes used glibenclamide plus metformin, miglitol, or placebo for 24 weeks in addition to their earlier therapy, fasting blood glucose concentrations improved with miglitol [ ]. Flatulence and diarrhea were significantly more common with miglitol. No patient stopped taking miglitol because of adverse effects.

General adverse effects and adverse reactions

Acarbose and miglitol have been reviewed [ , , ]. Their major adverse effects are flatulence, abdominal discomfort, diarrhea, and bloating, particularly at the start of therapy, which sometimes prevent further use. They should not be given to patients with intestinal obstruction, malabsorption, inflammatory bowel disease, or hepatic impairment.