Diabetology International

, Volume 1, Issue 1, pp 35–41

Efficacy and adverse effects of low-dose nateglinide in early type 2 diabetes: comparison with acarbose in a crossover study

  • Toyoyoshi Uchida
  • Junko Kawai
  • Yoshio Fujitani
  • Ryuzo Kawamori
  • Hirotaka Watada
  • Takahisa Hirose
Original article

DOI: 10.1007/s13340-010-0002-y

Cite this article as:
Uchida, T., Kawai, J., Fujitani, Y. et al. Diabetol Int (2010) 1: 35. doi:10.1007/s13340-010-0002-y
  • 84 Downloads

Abstract

Recent studies have suggested that treatment of type 2 diabetes soon after diagnosis of diabetes correlates with a long-term reduction in the risk of macroangiopathy. An open-label prospective crossover trial was performed to compare the efficacy and adverse effects of low-dose nateglinide with those of acarbose. The subjects were 23 early type 2 Japanese diabetic patients who were oral hypoglycemic agent naïve and whose HbA1C levels were 6.0–6.9%. Twelve patients received 90 mg/day of nateglinide and 11 patients received 150 mg/day of acarbose. After 12 weeks of either therapy, the drugs were switched and treatment was continued for another 12 weeks. Each patient was requested to complete a questionnaire on complications after the completion of each treatment. HbA1C decreased from a mean baseline of 6.7 ± 0.2 to 6.4 ± 0.3% with nateglinide and to 6.5 ± 0.3% with acarbose. Decreases of HbA1C in nateglinide were larger than those in acarbose, albeit statistically insignificant (p = 0.073). Symptoms related to hypoglycemia were scarcely observed with either treatment. Abdominal fullness/borborygmi were frequently reported by patients on acarbose but were absent or mild in those on nateglinide. After the completion of the study, most patients favored nateglinide over acarbose. Our results suggest that the effects of low-dose nateglinide on postprandial hyperglycemia of early type 2 diabetes are equivalent to those of acarbose, with the added advantage of minimal side effects.

Keywords

Postprandial hyperglycemia Glinide Alpha-glucosidase inhibitor Hypoglycemia Adverse effects 

Introduction

Recent progress in clinical research indicates that postprandial hyperglycemia is a risk factor for both micro- and macroangiopathy [1, 2], and the efficacy of the treatment of this abnormality has been reported or is currently under investigation [3, 4, 5, 6, 7]. Thus, there is a general agreement that patients with early type 2 diabetes should be treated more intensively and that efforts should be directed towards developing effective and acceptable oral hyperglycemic agents. We reported previously that a highly physiologic mealtime glucose regulator, nateglinide, was an effective and safe drug for the treatment of early type 2 Japanese diabetics with HbA1C levels between 7.0 and 7.9%, relative to the alpha-glucosidase inhibitor voglibose [8]. However, postprandial hyperglycemia is even observed in type 2 diabetes patients with HbA1C levels of less than 7.0%, indicating the need for intervention and improvement.

Nateglinide selectively enhances early meal-induced insulin secretion and thus improves postprandial hyperglycemia [9, 10, 11, 12]. The major adverse effect of nateglinide is hypoglycemia, and thus its use has been limited to patients with advanced type 2 diabetes on alpha-glucosidase inhibitors rather than those with early type 2 diabetes. A smaller dose of nateglinide may permit its use in patients with very early type 2 diabetes.

The aim of the present crossover study was to compare the effects and safety of low-dose (30 mg per meal, 90 mg/day) nateglinide for early type 2 diabetes with those of acarbose. The latter is an alpha-glucosidase inhibitor that has been proven in numerous clinical trials to be efficacious when administered alone or in combination with other antidiabetic agents in type 2 diabetic patients [13], and is reported to have sustained efficacy in long-term studies [14, 15].

Materials and methods

Study subjects and study design

This randomized open-label crossover study was conducted at Juntendo University Hospital and Chiba Tokushukai Hospital in Japan from January 2006 to January 2007. Twenty-three type 2 diabetic outpatients under diet and exercise therapy were randomly selected and enrolled in this study. Patients eligible for enrollment were oral hypoglycemic agent naïve, and had HbA1C of between 6.0 and 6.9% for more than 3 months before the start of this study. At the time of enrollment in the study, patients were divided randomly into two groups. The first group was started on 90 mg/day of nateglinide (30 mg before each meal) (n = 12), while the second group was started on 150 mg/day of acarbose (50 mg before each meal) (n = 11). A postprandial blood sample 2 h after regular breakfast was collected from each patient at baseline (before commencement of therapy). Regular breakfast was defined as the one that was the most typical style of breakfast for each patient. We requested the patients not to change the contents of this breakfast at baseline and after the completion of each drug instead of using a standard meal. They were treated for 12 weeks and subsequently switched to the opposite drug [nateglinide to acarbose (N–A group) and acarbose to nateglinide (A–N group)] for 12 weeks, with the collection of more blood samples. A questionnaire was provided to each patient after the completion of each treatment protocol on the adverse effects encountered during the treatment, such as hypoglycemic symptoms, change of appetite, and abdominal fullness or borborygmi. This included questions on the severity of these symptoms. Patients had to choose a number from 0 (no symptoms) to 10 (the most serious symptom). After the completion of both treatments, another questionnaire was provided which asked the patients to choose their favored therapy from among the two test antidiabetic agents (denoted the “favorite score”). This questionnaire also consisted of a visual scale from a5 (most favorable to acarbose) to n5 (most favorable to nateglinide), and each patient chose one number.

During this study, none of the patients were prescribed any other hypoglycemic agent. Moreover, none of them were on antihypertensive, hypolipidemic, or antiplatelet drugs during the study. Patients were excluded based on the following criteria: presence of concomitant chronic disease, including anemia, kidney, liver, and cardiovascular disease; and recent acute illness. All subjects were instructed not to change their usual dietary habits for the duration of the study. Informed consent was obtained from each subject. They were informed that their privacy would be protected and that their personal information would remain confidential. The study was well-conducted ethically.

Statistical analysis

Data are expressed as mean ± SD. Laboratory data were compared between the two groups using Student’s paired t test. Differences in HbA1C, postprandial plasma glucose level, and changes in body weight between the two drugs were analyzed using repeated measures ANOVA. The Friedman test was used to compare the severities of the adverse effects of the two drugs. Spearman’s correlation coefficient was used to check for factors that influenced glycemic control among the favorite score and adverse effects. A p value of less than 0.05 was taken to denote a significant difference between the groups.

Results

The demographic characteristics and baseline data of the patients of the two groups are summarized in Table 1. All subjects who completed the study reported taking their medications almost exactly as specified. The baseline laboratory data in the postprandial state are also shown in Table 1 for both groups. Age, gender, duration of diabetes, blood pressure, body mass index, HbA1C, and lipid profile (triglyceride, HDL-cholesterol, LDL-cholesterol) at baseline were not significantly different between the two groups.
Table 1

Baseline characteristics of the subjects

 

A–N group

N–A group

p value

All

n

11

12

 

23

Age (years)

63.0 ± 11.0

62.5 ± 15.0

0.936

62.8 ± 12.8

Sex (M/F)

7/4

5/7

 

12/11

Known duration of diabetes (years)

5.9 ± 2.7

4.8 ± 3.2

0.363

5.3 ± 3.0

sBP (mmHg)

124.1 ± 14.0

123.7 ± 14.1

0.95

123.9 ± 13.7

dBP (mmHg)

71.0 ± 7.0

70.9 ± 10.2

0.979

71.0 ± 8.5

T-Cho (mg/dl)

212.0 ± 22.5

198.8 ± 34.8

0.334

205.7 ± 28.9

HDL (mg/dl)

57.6 ± 18.0

56.9 ± 17.3

0.924

57.3 ± 17.2

TG (mg/dl)

156.4 ± 81.9

136.0 ± 42.4

0.472

146.2 ± 64.5

LDL (mg/dl)

119.7 ± 22.7

124.1 ± 18.4

0.748

122.8 ± 18.6

HbA1C (%)

6.8 ± 0.2

6.7 ± 0.2

0.641

6.7 ± 0.2

BMI

23.6 ± 2.7

24.6 ± 5.6

0.583

24.1 ± 4.3

Mean ± SD values are shown

Efficacy of glycemic control

The parameters used to evaluate each drug are summarized in Table 2. The primary efficacy parameter used in this study was HbA1C. After 3 months of treatment with each drug, HbA1C decreased from a baseline value of 6.7 ± 0.2 to 6.4 ± 0.3% with nateglinide and to 6.5 ± 0.3% with acarbose (nateglinide vs. acarbose, A–N group, n = 11, 6.4 ± 0.3 vs. 6.6 ± 0.3%, N–A group, n = 12, 6.4 ± 0.3 vs. 6.5 ± 0.2%). A comparison of the two drugs showed that nateglinide tended to produce a larger decrease in glycosylated hemoglobin than acarbose, but the difference was not statistically significant (p = 0.073).
Table 2

Results of primary parameters used to evaluate nateglinide and acarbose

 

n

Baseline

Nateglinide

Acarbose

p valuea

Laboratory data

 HbA1C

23

6.7 (0.2)

6.4 (0.3)

6.5 (0.3)

0.0729

 2 h PPG

23

171.0 (37.0)

136.2 (33.8)

150.3 (33.4)

0.0058

 Weight

23

60.0 (13.4)

59.7 (13.3)

59.9 (13.2)

0.6018

Mean (SD) values are shown

aRepeated measures ANOVA was employed (with patients as a random effect and treatment and order of treatment as fixed effects)

The other efficacy parameter measured was postmeal glucose level. The mean level of postmeal blood glucose (2 h after regular breakfast) decreased from a baseline value of 171.0 ± 37.0 to 136.2 ± 33.8 mg/dl with nateglinide and to 150.3 ± 33.4 mg/dl with acarbose, and the effect of nateglinide on the postmeal glucose level was significantly greater than that of acarbose (p = 0.0058).

The mean body weight after the use of either drug was not different from the baseline or different from that measured after the use of the other agent (nateglinide vs. acarbose, baseline 60.0 ± 13.4 kg, nateglinide 59.7 ± 13.3 kg vs. acarbose 59.9 ± 13.2 kg, p = 0.60). A comparison of other biochemical data measured before and after each drug also showed no significant difference, including serum lipid levels (data not shown). Based on these results, we concluded that low-dose nateglinide was a suitable and effective drug for treating early type 2 diabetes, similar to acarbose.

Adverse effects

The major concern with acarbose treatment is abdominal fullness/borborygmi, while that with nateglinide is hypoglycemia. In this study, all patients were asked to report the severity of gastrointestinal symptoms, including abdominal fullness and borborygmi, at each follow-up visit. Table 3 lists the mean severity scale for adverse effects (0––none, 1––least serious, 10––most serious) during either treatment. During acarbose treatment, 21 of the 23 patients reported gastrointestinal symptoms, including borborygmi (21/23, mean severity score 6.7) and abdominal fullness (19/23, mean severity score 4.3), while only 1 or 2 of those on nateglinide reported such symptoms, and the severity was obviously lower (Table 3).
Table 3

Severity and frequency of adverse effects of nateglinide and acarbose

 

n

Nateglinide

Acarbose

p valuea

AE score

 Abdominal fullness

23

0.3

(0–4, 2)

4.3

(0–8, 19)

<0.0001

 Borborygmi

23

0.1

(0–3, 1)

6.7

(0–10, 21)

<0.0001

 Abdominal pain

23

0.2

(0–3, 2)

0.3

(0–3, 3)

0.6400

 Constipation/diarrhea

23

0.5

(0–7, 3)

0.8

(0–4, 6)

0.2736

 Increased appetite

23

0.8

(0–5, 7)

0.3

(0–3, 3)

0.1435

 Shivering

23

0.0

0.0

 Sweating

23

0.0

0.0

 Palpitation

23

0.0

0.0

Mean (range, number of patients for whom the severity score was more than 1)

aFriedman test (order of treatment as a strata)

The patients were also asked about hypoglycemia-related symptoms at each follow-up visit. These symptoms included increased appetite, shivering, sweating, and palpitation. After nateglinide or acarbose, the patients showed no hypoglycemic symptoms, with the exception of increased appetite. During nateglinide treatment, 7 of the 23 patients reported a mild increase in appetite, but the mean severity score was low (mean severity score 0.8, range 0–5), while during acarbose, 4 patients reported the same (mean severity score 0.3, range 0–3). Neither drug induced severe hypoglycemia, even though the level of glycosylated hemoglobin was between 6.0 and 6.9%.

Favorite score

After the completion of both treatments, another questionnaire requested information about favored therapy from among the two test antidiabetic agents. This questionnaire also consisted of a scale from a5 (perfectly favorable to acarbose) to n5 (perfectly favorable to nateglinide). Twenty of the 23 patients selected a scale number that was more than a score of 1. The mean favorable score was n3.5 ± 2.3 (p < 0.0001), suggesting that most of the patients favored nateglinide over acarbose.

The relationship between glycemic control and adverse effects was examined by Spearman’s correlation coefficient analysis. Abdominal fullness correlated with HbA1C during the administration of acarbose (r = 0.59, p = 0.0032, Table 4). Meanwhile, the adverse effects did not correlate with HbA1C during the administration of nateglinide. These results suggest that severe gastrointestinal adverse effects induced by acarbose could be the reason for the reduced compliance, even though these patients reported complete intake of the drugs.
Table 4

Correlation between favorite score or adverse effects and HbA1C

 

n

Nateglinide

Acarbose

Corr.

p value

Corr.

p value

Spearman’s correlation coefficient (favorite score or AEs vs. HbA1C)

 Favorite score

23

−0.39

0.0679

0.10

0.6523

 Abdominal fullness

23

0.15

0.4841

0.59

0.0032

 Borborygmi

23

0.22

0.3153

0.23

0.2881

 Abdominal pain

23

−0.10

0.6408

0.20

0.3557

 Constipation/diarrhea

23

−0.35

0.1008

0.19

0.3731

 Increased appetite

23

0.04

0.8478

0.24

0.2656

 Tremor

23

 Sweating

23

 Palpitation

23

Next, we investigated the relationship between favorite score and glycemic control after treatment with either of the drugs (Table 4). The favorite score tended to correlate negatively with HbA1C during the administration of nateglinide, albeit insignificantly (r = −0.39, p = 0.0679). These results suggest that patients may elect to use nateglinide based on its effectiveness and minimal adverse effects.

The results suggest that low-dose nateglinide (90 mg/day) is as effective as acarbose (150 mg/day) for glycemic control in patients with early type 2 diabetes, and it seems to be more acceptable based on its minimal adverse effects.

Discussion

One goal of the treatment of nonpregnant adults is to lower HbA1C to less than 7.0% in order to prevent microvascular and neuropathic complications of type 1 and type 2 diabetes [16]. Randomized controlled trials of intensive versus standard glycemic control have shown no significant changes in the outcome of cardiovascular disease (CVD) during the randomized portion of the trials. However, long-term follow-up of the Diabetes Control and Complication Trial (DCCT) [17] and the UK Prospective Diabetes Study (UKPDS) [18] cohorts suggests that treatment to HbA1C targets below or around 7.0% soon after the diagnosis of diabetes is associated with a long-term reduction in the risk of macroangiopathies. Other recent randomized controlled trials like ADVANCE [19] and the Action to Control Cardiovascular Risk in Diabetes Study Group (ACCORD) [20] have reported that intensive glycemic control (HbA1C to around 6.5%) did not succeed in reducing the risk of macrovascular disease. The mean duration of known diabetes for the patients who participated in these studies was around 10 years. VADT [21] also did not show any reduction in the risk of CVD with intensive glycemic control (to around 6.5%), but the risk hazard ratio of CVD in that control correlated with the duration of diabetes in comparison with the standard control in this subanalysis. Although the American Diabetes Association (ADA) recommendation for HbA1C < 7.0% appears reasonable in many adults for macrovascular risk reduction until more evidence becomes available [16], initiating intervention after the diagnosis of diabetes sooner may lead to a lower risk of CVD. How soon the treatment of diabetes is initiated may be very important in addition to the level of the glycemic goal.

In this sense, one of the major targets of glycemic control in early type 2 diabetes is postprandial hyperglycemia. Postprandial hyperglycemia observed even in individuals with impaired glucose tolerance may be a determinant of future CVD, and the efficacy of intervention to treat this abnormality has been reported or is currently under investigation [1, 2]. We reported previously that nateglinide reduced carotid intima-media thickening in type 2 diabetic patients under good glycemic control (HbA1C < 6.5%) [5]. Recently, Holman et al. reported in the NAVIGATOR study that assignment to nateglinide for 5 years did not reduce the incidence of diabetes or coprimary composite cardiovascular outcomes among persons with impaired glucose tolerance and established cardiovascular disease or cardiovascular risk factors [22]. These negative results, in particular for the composite cardiovascular outcomes, may be due to the short duration of the intervention period in comparison with other positive studies like UKPDS 80 [18]. The STOP-NIDDM trial has reported a significant risk reduction of cardiovascular events, in particular myocardial infarction, in an acarbose group after almost the same duration as in the NAVIGATOR study. However, it is important to note that the number of myocardial infarctions was very low (acarbose: 1; control: 12) in the STOP-NIDDM trial. In this trial, there was no significant difference between acarbose and placebo for several cardiovascular events with primary endpoints, except for myocardial infarction. The above background suggests that there is a need for more intensive treatment of patients with early type 2 diabetes and to search for not only effective but also acceptable oral hyperglycemia agents for very early type 2 diabetes.

We reported previously that the highly physiologic mealtime glucose regulator nateglinide was an effective and safe drug for the treatment of Japanese early type 2 diabetics with HbA1C levels of between 7.0 and 7.9%, compared with the alpha-glucosidase inhibitor voglibose [8]. In the present study, we showed that the administration of low-dose nateglinide and acarbose resulted in similar improvements in early type 2 diabetic patients whose glycosylated hemoglobin levels ranged from 6.0 to 6.9%. The primary parameter of efficacy measured was HbA1C. After 3 months of treatment with each drug, HbA1C decreased from a baseline of 6.7 ± 0.2 to 6.4 ± 0.3% with nateglinide and to 6.5 ± 0.3% with acarbose. A comparison of the two drugs showed that nateglinide caused a somewhat greater decrease in glycosylated hemoglobin than acarbose, but the difference was not statistically significant (p = 0.073).

The other efficacy parameter measured in the present study was postmeal glucose level. The mean level of postmeal plasma glucose decreased from a baseline level of 171.0 ± 37.0 to 136.2 ± 33.8 mg/dl with nateglinide and to 150.3 ± 33.4 mg/dl with acarbose, and the effect of nateglinide on postmeal glucose level was significantly stronger than that of acarbose (p = 0.0058). Reduction of postprandial glucose levels with nateglinide was significantly greater than that achieved with acarbose, but the reduction in HbA1c was not significantly different between the two treatments. The possible reasons for this discrepancy are the small sample size of this study and the limitations of HbA1c, which may not be a direct index for postprandial glucose levels.

At present, various therapies are available for directly improving postprandial hyperglycemia. One of these is to use alpha-glucosidase inhibitors, while another is to use “glinides,” which are rapid insulin secretagogues. Acarbose, miglitol and voglibose act by competitively inhibiting alpha-glucosidases, a group of key intestinal enzymes involved in the digestion of carbohydrates. They reduce both postprandial hyperglycemia and hyperinsulinemia [23]. These compounds, including acarbose, have been proven to be efficacious when administered alone or in combination with other antidiabetic agents in numerous clinical trials involving type 2 diabetic patients [13], and they have shown sustained efficacy in long-term studies [14, 15]. Therefore, these agents are widely prescribed for early type 2 diabetes patients, especially in Japan. However, the adverse gastrointestinal effects of these drugs may limit long-term compliance to therapy. In this study, nateglinide, one of the “glinides,” selectively enhanced early meal-induced insulin secretion and thus improved mealtime glucose control. We reported previously that nateglinide improved glycemic response after oral glucose load in obese individuals with impaired glucose tolerance or early type 2 diabetes by improving early phase insulin secretion without increasing the total amount of insulin (i.e., the area under the curve) [11, 12].

Only a few patients in the present study reported having symptoms related to hypoglycemia (increased appetite, palpitation, sweating, or shivering), and this was true for both nateglinide as well as acarbose. We have also compared the effects of a small dose of gliclazide (20 mg) and nateglinide (270 mg), and we reported that the hypoglycemic effects of glicalzide tended to be stronger though the drug elicited more frequent hypoglycemic episodes than nateglinide [24].

On the other hand, abdominal fullness/borborygmi were frequently reported in patients treated with acarbose, though they varied in severity [25], while no such complaints, or much milder episodes, were reported by those in the nateglinide group. The reported frequency and severity of gastrointestinal side effects of alpha-glucosidase inhibitors varies in different studies. Scorpiglione et al. [26] reported that ~33% of their type 2 diabetic patients could not continue on acarbose for the whole study period, mainly due to its gastrointestinal side effects. In a recent prevention study of a wide spectrum of the disease from impaired glucose tolerance to type 2 diabetes using acarbose, 13% of the acarbose group prematurely discontinued the therapy due to gastrointestinal symptoms, compared with 3% of the placebo group [27]. In the present study, the severity of gastrointestinal adverse effects induced by acarbose correlated with the level of HbA1C, suggesting the possibility that these unfavorable adverse effects decrease the compliance of this drug, even though all of the patients declared that they took acarbose with full compliance. The reason for the preference of nateglinide as the drug of choice in our study may be the lack of adverse gastrointestinal effects and the effectiveness of its glycemic control. To minimize the adverse gastrointestinal effects of acarbose, it is advised to use a “start low, go slow” policy. This titration of the drug was not performed in our present study, and it may have improved the favorite scores of acarbose somewhat. Our present study was performed as an open-label study. We could not eliminate the possibility of the results of the questionnaire being influenced in such an open-label study. We need to confirm our data by performing a placebo-controlled double-blind randomized study in the future.

There is little or no information on the differences in the effects of acarbose and nateglinide on other metabolic markers. A Chinese study involving a small number of patients reported that nateglinide acutely decreased postprandial free fatty acid and asymmetric dimethylarginine levels, and that these changes were significantly more profound than acarbose [28]. Our study did not analyze these parameters. Further efforts are needed to investigate these aspects of treatment and to identify the subjects who would respond best to each class of drugs.

In conclusion, our data indicate that low-dose nateglinide (90 mg/day) is as effective as acarbose (150 mg/day) for glycemic control in patients with early type 2 diabetes, though it seems to be a more acceptable oral hypoglycemic agent based on its fewer side effects.

Study limitations

Certain study limitations of the present study should be recognized. The study comprised a comparatively small number of patients and was conducted in only a few institutions. Treatment with nateglinide reduced postprandial glucose levels significantly compared with acarbose, but HbA1c did not show a significant difference between the two treatments. A possible reason for this discrepancy may be the small sample size of this study. In general, further studies are needed to evaluate the long-term effects of nateglinide and acarbose in relation to hypoglycemic effects and postprandial hyperglycemia that involve the inclusion of more patients after more intensive education.

Conflict of interest

This clinical trial received financial support from Daiichi Sankyo. Daiichi Sankyo did not have any influence on the analysis and interpretation of the data. T. Hirose, Y. Fujitani and R. Kawamori have received grant support from Takeda. RK, HW, and TH have also acted as spokespeople for Daiichi Sankyo. RK has received grant support from Nippon Eli Lilly. All of the other authors declare no conflict of interest.

Copyright information

© The Japan Diabetes Society 2010

Authors and Affiliations

  • Toyoyoshi Uchida
    • 1
  • Junko Kawai
    • 1
  • Yoshio Fujitani
    • 1
    • 2
  • Ryuzo Kawamori
    • 1
    • 2
    • 3
    • 4
  • Hirotaka Watada
    • 1
  • Takahisa Hirose
    • 1
    • 2
  1. 1.Department of Medicine, Metabolism and EndocrinologyJuntendo University Graduate School of MedicineTokyoJapan
  2. 2.Center for Therapeutic Innovations in DiabetesJuntendo University Graduate School of MedicineTokyoJapan
  3. 3.Center for Beta Cell Biology and RegenerationJuntendo University Graduate School of MedicineTokyoJapan
  4. 4.Sportology CenterJuntendo University Graduate School of MedicineTokyoJapan

Personalised recommendations