Low-Dose Acarbose Improves Glycemic Control in ... - Diabetes Care

S H O R T

R E P O R T

W

e have demonstrated previously that in healthy subjects, administering the a-glucosidase inhibitor, acarbose, as a powder mixed uniformly through the meal produced significant, dose-dependent reductions in postprandial glucose and insulin responses over the 12.5-50 mg dose range (1). The optimum result of maximum flattening of the postprandial metabolic ALISON NANKERVIS, MD, FRACP ADAM JENNEY, BMEDSCI, MBBS responses without detectable carbohyJOE PROIETTO, MBBS, PHD, FRACP KATHY TRAIANEDES, MAPPSC KERIN O'DEA, PHD HELEN D'EMBDEN, BSC, GRADDIPDIET drate malabsorption (as recorded by breath hydrogen measurements and subjective reporting of symptoms) was obOBJECTIVE — To examine the impact on metabolic control in NIDDM patients of tained with an acarbose dose of 25 mg. the a-glucosidase inhibitor, acarbose, when administered at a low dose in powdered The aim of this study was to determine, first, whether improvements in metaform. bolic control in NIDDM reported by othRESEARCH DESIGN AND METHODS — Six subjects were recruited for a dou- ers (2,3) who used acarbose in the tablet ble-blind cross-over trial using 25 mg powdered acarbose and a placebo 3 times a form at high doses (100-200 mg/meal) day with meals for 3 mo. In addition to parameters of diabetes control and body could be demonstrated when it was adweight, glucose turnover and insulin sensitivity were measured with the hyperinsu- ministered at a much lower dose (25 linemic/euglycemic clamp technique combined with tracer kinetics. mg/meal) in the powdered form; and RESULTS — None of the subjects showed significant changes in FPG levels or body second, whether the symptoms of carboweight either on the 3-mo course of acarbose or placebo. HbAlc fell significantly from hydrate malabsorption commonly re10.6 ± 1.0 to 9.4 ± 1.3% (P = 0.05) during treatment with acarbose but failed to ported with acarbose therapy, such as change on placebo (10.1 ± 1.0 to 11.1 ± 2.0%; P = 0.36). Basal HGP and glucose flatulence, colic, and meteorism (3-5), utilization were unchanged during either of the treatment periods, and hyperinsu- could be attenuated by this different linemia produced a similar degree of suppression of HGP before and after each mode of acarbose administration. To this treatment. At a physiological concentration, insulin failed to stimulate glucose clear- end, 6 NIDDM patients were recruited ance in these diabetic patients, and no improvement was seen with acarbose treat- for a double-blind crossover trial using ment. No changes in plasma lipids or lipoprotein profiles were demonstrated after 3 25 mg powdered acarbose or placebo 3 mo on acarbose. In acute studies, it was shown that administration of acarbose at a time/day with meals for 3 mo. Metabolic dose of 25 mg powder per meal significantly decreased the postprandial glycemic control, glucose turnover, and insulin sensitivity were measured before and afexcursion. ter each 3-mo treatment period (acarbose CONCLUSIONS — When administered in the powdered form at the low dose of or placebo). 25 mg 3 times/day with meals over 3 mo, acarbose was well tolerated by the NIDDM patients and was without side effects. It improved glycemic control by reducing postprandial hyperglycemia, but had no effect on glucose turnover, insulin sensitiv- RESEARCH DESIGN AND METHODS— The study protocol was ity, or lipid profile. approved by the Royal Melbourne Hospital Ethics Committee. Six nonobese NIDDM patients (4 men, 2 women) were FROM THE DEPARTMENT OF MEDICINE, UNIVERSITY OF MELBOURNE, ROYAL MELBOURNE HOSPITAL, recruited from the outpatient clinic of MELBOURNE; AND THE DEPARTMENT OF HUMAN NUTRITION, DEAKIN UNIVERSITY, GEELONG, VICTORIA, the Royal Melbourne Hospital. Their AUSTRALIA. mean age was 60.3 ± 2.5 yr, and mean ADDRESS CORRESPONDENCE AND REPRINT REQUESTS TO KERIN O ' D E A , PHD, DEPARTMENT OF HUMAN body weight was 72.3 ± 2.9 kg. They NUTRITION, DEAKIN UNIVERSITY, GEELONG, VICTORIA, AUSTRALIA 3 2 1 7 . were controlled by oral hypoglycemic RECEIVED FOR PUBLICATION 18 DECEMBER 1991 AND ACCEPTED IN REVISED FORM 2 8 OCTOBER 1992. N I D D M , NON-INSULIN-DEPENDENT DIABETES MELLITUS; F P G , FASTING PLASMA GLUCOSE; H G P , HEagents or diet alone. They underwent a PATIC GLUCOSE PRODUCTION; RJ,, RATE OF GLUCOSE DISPOSAL; AUC, AREA UNDER THE CURVE. medical examination, which included investigations of renal and hepatic func-

Low-Dose Acarbose Improves Glycemic Control in NIDDM Patients Without Changes in Insulin Sensitivity

DIABETES CARE, VOLUME 16, NUMBER 2, FEBRUARY

1993

499

Acarbose and glycemic control in diabetes

Table 1—Body weight and metabolic parameters measured in the fasting state throughout the study ACARBOSE TREATMENT

PLACEBO TREATMENT

BEFORE BODY WEIGHT (KG) HBAU

(%)

CHOLESTEROL (MM) TRIGLYCERIDE (MM)

71.6 10.2 5.6 2.0

±3.1 ± 1.2 ± 0.5 ± 0.3

AFTER

P VALUE

BEFORE

72.2 ± 2.9 11.1 ±2.0 5.2 ± 0.6 1.6 ±0.3

0.26 0.36 0.21 0.40

71.2 ±3.0 10.6 ± 1.0 5.3 ± 0.5 2.5 ± 0.7

AFTER

70.7 9.4 5.3 2.3

±3.1 ± 1.3 ± 0.7 ± 0.3

P VALUE

0.24 0.05 0.34 0.86

Data are means ± SE, n = 6.

tion. A dietitian analyzed their individual diet by using a 7-day weighed-food record and the MicroDiet software package based on the English food tables (5). During the 2-mo run-in period, their diets were modified, if necessary, to approach the recommended dietary treatment for N1DDM: 30% energy from fat and 50-60% energy from carbohydrate. For the study of a drug that acts by delaying carbohydrate digestion and absorption, a high-carbohydrate diet was considered desirable. In terms of reduction of cardiovascular complications of diabetes, a low-fat diet is recommended. Thus, the patients' long-term dietary therapy was not compromised by participation in this study. Subjects underwent weekly dietary assessment, monthly physical examination, and 6 weekly breath hydrogen measurements during the two 3-mo trial periods (acarbose or placebo). At the beginning and end of each 3-mo period, basal glucose turnover (7), insulin sensitivity, cholesterol and triglyceride levels, Hb and white cell count, HbAlc, and urea and electrolytes were measured; and liver function tests were performed. There was a 1-mo break between the two trial periods, with 5 of 6 subjects completing the acarbose period first. To confirm that the dose of acarbose being administered chronically in the long-term study was producing the expected effects on postprandial glucose and insulin responses to a carbohydrate load acutely, 5 of 6 subjects agreed to be tested on two occasions after the com-

500

pletion of the long-term study. On the mornings of the acute studies, a fasting blood sample was taken, and the subjects then consumed a standard breakfast containing 75 g carbohydrate (cornflakes, milk, orange juice, toast) onto which 25 mg acarbose powder or placebo had been sprinkled. The meal was consumed within 15 min, and blood samples taken over the next 2 h were analyzed for glucose and insulin.

Analytical procedures Glucose, insulin, lipids, percentage of enrichment of the dideuterated glucose, and breath hydrogen were analyzed by standard methods as described previously (1,8). Calculations The parameters of glucose turnover were calculated as described previously (7,8). The paired Student's t test was used to compare the effects of acarbose versus placebo on the measured parameters. RESULTS— No significant changes were observed in energy intake or dietary composition over time, nor were any differences noted in energy intake or dietary composition between the two dietary periods. Consistent with this, body weight remained constant throughout the 6 mo of the study. Breath hydrogen measurements taken throughout the study showed no significant effect of acarbose or placebo on hydrogen production,

which is consistent with no significant carbohydrate malabsorption during either phase of the study. Although HbAlc did not change significantly when the patients were on the placebo, this value fell significantly after acarbose. The decrease in HbAlc was modest, but it occurred in all subjects. No differences were seen in fasting cholesterol or triglyceride levels (Table 1). None of the parameters of glucose turnover in the basal state or during the hyperglycemic clamp (basal glucose or insulin levels, basal HGP, or Rj were changed by acarbose treatment (Table 2). During insulin infusion, insulin levels were increased to the low physiological range, and glycemia was maintained at fasting levels. As can be seen, this amount of insulin resulted in suppression of hepatic glucose output to < 50% of the basal value. However, this amount of insulin was insufficient to stimulate glucose clearance above basal. Acarbose treatment did not alter insulin sensitivity to either inhibition of HGP or stimulation of Rj. Figure 1 shows the glucose (Fig. 1A) and insulin (Fig. IB) from the acute meal test on the study group. An amount of 25 mg acarbose sprinkled over the meal significantly reduced the glycemic excursion (AUC glucose acarbose 1890 ± 186 m M / m i n ; p l a c e b o 2092 ± 204 mM/min, P < 0.05), and the trend was towards lower insulin levels.

DIABETES CARE, VOLUME 16,

NUMBER 2,

FEBRUARY

1993

Jenney and Associates

Table 2—Glucose and insulin concentrations, HGP, and Rd in the fasting state and during the hyperglycemic clamp studies conducted before and after the acarbose and placebo treatment periods PLACEBO TREATMENT BEFORE

AFTER

ACARBOSE TREATMENT P VALUE

BEFORE

AFTER

P vAi.ui-:

FASTING STATE GLUCOSE (MM) INSULIN (MU/L) H G P ((XMOL • KG" 1 • MIN" 1 ) R J ((XMOl. • KG" 1 • MIN" 1 )

9.8 6.8 31.3 33.9

± 1.7 ± 0.6 ±6.0 ± 6.2

10.8 8.3 27.6 27.2

± ± ± ±

1.7 1.4 5.4 5.3

0.14 0.24 0.51 0.50

9.4 6.7 24.5 25.1

± ± ± ±

1.5 1.1 5.5 5.4

9.6 7.4 23.6 23.6

± ± ± ±

1.1 0.5 4.3 4.3

0.38 0.55 0.18 0.09

10.0 37.0 13.4 26.2

± 1.1 ± 1.9 ±2.8 ± 4.0

9.7 36.6 9.4 27.9

± ± ± ±

1.0 4.2 4.1 4.8

0.2 0.93 0.36 0.85

8.5 37.4 11.5 31.7

± 1.4 ± 4.8 ±2.5 ±5.3

9.9 40.2 7.8 26.1

± ± ± ±

1.1 4.7 0.9 4.6

0.13 0.74 0.10 0.28

HYPERGLYCEMIC CLAMP GLUCOSE (MM) INSULIN (MU/L) HGP

(JJIMOL • K G " 1 • M I N " 1 )

R j (jJLMOL- KG" 1 • MIN" 1 )

Data are means ± SE, n = 6.

A

o

Time (min)

~n 30

60

90

120

Time (min)

Figure 1—Change in glucose (A) and insulin (B) levels in response to the standard breakfast in the presence of acarbose (—) (mean ± SE, n = 5).

or placebo (

)

CONCLUSIONS— The major finding in this study was that when acarbose was administered as a powder sprinkled on food 3 times/day with meals, it resulted in improved glycemic control in NIDDM patients, when used at a low dose with no detectable side effects related to carbohydrate malabsorption. No changes were seen with either the placebo or acarbose in fasting blood glucose, insulin, basal glucose turnover, or insulin sensitivity. However, a small but statistically significant fall was observed in HbAlc levels (P = 0.05 on a two-tailed paired Student's t test) after 3 mo on acarbose. In contrast, a 1% increase was seen in mean HbAlc during the placebo period, although this change was not statistically significant. This modest fall in HbAlc was consistent with a reduction in postprandial glycemic excursion noted in the acute meal studies. Basal HGP varied widely between patients but was closely correlated with fasting glycemia (r = 0.897). Acarbose had no effect in lowering basal HGP or the basal rate of glucose disappearance. Fasting insulin levels did not change during the study. The insulin infused at 40 mU • kg" 1 • h" 1 resulted in plasma insulin concentrations of —40 mU/L. This insulin level was sufficient to partially suppress HGP. However, no appar-

DIABETES CARE, VOLUME 16, NUMBER 2, FEBRUARY 1993

ent effect of acarbose to enhance this suppressant action of insulin was seen. Although the insulin dose infused was sufficient to partially suppress HGP, it was insufficient to alter peripheral glucose uptake, as evidenced by the failure of R^j to increase. This is not surprising because suppression of HGP can be demonstrated at lower insulin levels than those needed to stimulate peripheral glucose uptake in nondiabetic subjects (9). The relatively low insulin infusion level in this study was chosen as one which would reveal even subtle effects of acarbose on either hepatic or peripheral insulin sensitivity were they present. Although at higher doses acarbose has been reported to produce a small improvement in hepatic insulin sensitivity (3), it has not been shown to improve peripheral insulin action (10). Despite this, numerous studies have reported improved glycemic control. On the basis of this study, it seems reasonable to conclude, in agreement with others (2), that any long-term benefits associated with the use of acarbose in the therapy of NIDDM patients (even at a much higher dose (2) than that used in this study) are attributable to its acute action in delaying carbohydrate absorption. The important observation from this study is that these beneficial effects

501

Acarbose and glycemic control in diabetes

can be achieved with a much lower dose of acarbose than is generally recommended, and one that is not associated with any significant gastrointestinal side effects, if the drug is administered as a powder distributed uniformly through the meal. Acknowledgments—This work was supported by grants from Bayer Australia Limited and the National Health and Medical Research Council of Australia. J.P. was a Wellcome Australia Senior Research Fellow.

References 1. O'Dea K, Turton J: Optimum effectiveness of intestinal a-glucosidase inhibitors: importance of uniform distribution through a meal. Am] Clin Nutr 41:51116, 1985 2. Reaven GM, Lardinois CK, Greenfield

502

MB, Schwartz HC, Vreman HJ: Effect of 7. De Fronzo RA, Tobin JD, Andres R: GluAcarbose on carbohydrate and lipid mecose clamp technique: a method for tabolism in NIDDM patients poorly conquantifying insulin secretion and resistrolled by sulfonylureas. Diabetes Care 13 tance. Am J Physiol 237:E214-23, (Suppl. 3):32-36, 1990 1979 Baron AD, Eckel RH, Schmeiser L, Kol8. Proietto J, Nankervis AJ, Traianedes K, terman OG: The effect of short-term alRosella G, O'Dea K: Identification of pha glucosidase inhibition on carbohyearly metabolic defects in young subjects drate and lipid metabolism in type II from a diabetes-prone Australian Aborig(non-insulin-dependent) diabetics. Meinal community. Diabetes Res Clin Pract tabolism 36:409-15, 1987 17:217-22, 1992 Caspary WF: Sucrose malabsorption in 9. Proietto J, Alford FP, Dudley FG: The man after ingestion of a-glucosidase inmechanism of carbohydrate intolerance hibitor. Lancet 1:1231-33, 1978 of cirrhosis. J Clin Endocrinol Metab 51: Scott RS, Knowles RL, Beaven DW: 1030-36, 1980 Treatment of poorly controlled non10. Buchanan DR, Collier SA, Caspary WF, insulin dependent diabetic patients with Rodrigues E, Millar AM, Gray RS, Clarke Acarbose. Aust NZ ] Med 14:649-54, BF: Effectiveness of acarbose, an alpha1984 glucosidase inhibitor, in uncontrolled Paul AA, Southgate DAT: McCance and non-obese noninsulin dependent diabeWiddowson's The Composition of Foods. tes. Eur J Clin Pharmacol 34:51-53, 4th ed. London, Her Majesty's Stationery 1988 Office, 1978

DIABETES CARE, VOLUME 16,

NUMBER 2,

FEBRUARY

1993