Diabetes Research and Clinical Practice - Volume 120 - Supplement 1

control in Taiwan. One hundred and sixty type 2 diabetic

patients were recruited from China Medical University

Hospital in Taichung, Taiwan. Biochemical data, such as

plasma fasting sugar, HbA1c, total cholesterol, triglyceride,

HDL-C, and LDL-C, were collected from medical records. Diet

was assessed by a validated food frequency questionnaire

which composed of 118 food items, and categorized to 23 food

groups. Diabetic patient

’

s dietary pattern was analyzed by

principle component analysis. Confounders related to both

diet and biochemical indices including age, gender, BMI,

energy intakes, and physical activity were assessed. Multiple

logistic regression analyses were used to predict odds ratios of

poor blood biochemical control after adjustment of confoun-

ders. Mean age of the study subjects was 57.46 ± 10.67 year-old

and had 2.8 year of diabetes history. Four dietary patterns were

identified, pattern I (high meat, fish, vegetable, nut and oil),

pattern 2 (high refined rice, starchy root, and fish), pattern 3

(high refined rice, meat product, sugar drink, low whole grain,

vegetable, nut), and pattern 4 (high vegetable, non-sugar drink,

low starchy root, dairy product, and snack). Mean age, energy

intake, BMI, and sex proportion were significantly different

among different patterns. Compared to pattern 4, the odds

ratio of high HbA1c (

≥

7%) was 3.66 (95% CI 1.21

–

11.05) for

pattern 1 and 2.52 (95% CI 1.00

–

6.39) for pattern 3 after gender,

age, energy intake, and BMI adjustment. Odds ratio of high

plasma fasting sugar (>130 mg/dL) was 2.67 (95% CI 1.08

–

6.65)

for pattern 3 compared to pattern 4. For female patients, odds

ratio of low HDL-C (<50 mg/dL) was 0.11 (95% CI 0.02

–

0.80) for

pattern 1 compared to pattern 4. In the present study, clusters

of food groups might not be representative for Taiwanese

diabetic patients; however, different dietary patterns were

possible to influence blood sugar and lipid control among

diabetic patients. More consumption of vegetables, fish and

nuts, less sugar drinks and snacks are beneficial for diabetic

treatment.

PD-94

Relationship between indices of daily glycemic variability and

long-term glucose fluctuations

Yuusuke WATANABE

*, Yoshifumi SAISHO

, Rie MURAKAMI

Jun INAISHI

, Kenji KANEKO

, Hiroshi ITOH

Department of

Internal Medicine, Keio University School of Medicine, Japan

Purpose:

We investigated the relationship between indices of

daily glycemic variability and long-term glucose fluctuations.

Methods:

A total of 123 patients with diabetes (22 patients with

type 1 diabetes; 9 males, age 53 ± 16, HbA1c 7.9 ± 1.4%, GA

25.0 ± 6.0%, and 101 patientswith type 2 diabetes; 80males, age

69 ± 10, HbA1c 6.8 ± 0.8%, GA 18.8 ± 3.9%) who underwent

continuous glucose monitoring (CGM) in outpatient settings

were enrolled in this study. We evaluated the daily glycemic

variability from the CGM data as follows; mean glucose level,

standard deviation (SD), time of hyper- and hypoglycemia,

area under the curve (AUC) of hyper- and hypoglycemia, J-

index, hyperglycemic index, hypoglycemic index, index of

glycemic control (ICG), continuous overlapping net glycemic

action (CONGA), mean amplitude of glycemic excursion

(MAGE), M-value, mean of the daily differences (MODD). In

addition, Coefficient of variation (CV) of HbA1c and GA were

calculated from HbA1c and GA measured before and after 3

years from CGM measurement.

Result:

Both CV-HbA1c and CV-GAwere correlated with mean

glucose level, SD, time of hyperglycemia, AUC of hypergly-

cemia, J-index, hyperglycemic index, ICG, MODD, MAGE and

M-value but were not correlated with time of hypoglycemia,

AUC of hypoglycemia, hypoglycemic index, and CONGA-1.

Especially, CV-HbA1c was strongly correlated with mean

glucose level, time of hyperglycemia, AUC of hyperglycemia,

J-index and M-value (r = 0.4), and so was CV-GA. Serum C-

peptide level was negatively correlated with the indices of

daily glycemic variability, CV-HbA1c and CV-GA, and was also

correlated with GA/HbA1c. GA/HbA1c was correlated with the

indices of daily glycemic variability, but not correlated with

CV-HbA1c or CV-GA. These results were not changed when

only the patients with type 2 diabetes were included in the

analysis.

Conclusion:

The indices of daily glycemic variability were

correlated with long-term glucose fluctuations, but were not

correlated with the indices of hypoglycemia. These findings

suggest that CV-HbA1c and CV-GA may not reflect the risk of

hypoglycemia. And

-cell function was correlated with both

daily glycemic variability and long-term glucose fluctuations.

Clarifying the features of various indices of glycemic variabil-

ity and the relationship among them couldmake these indices

more useful in clinical practice.

PD-95

Efficacy and safety of once-weekly semaglutide monotherapy

versus placebo in subjects with type 2 diabetes (SUSTAIN 1)

Shin-ichi HARASHIMA

*, Christopher SORLI

George TSOUKAS

, Jeffrey UNGER

, Julie DERVING KARSBØL

Thomas HANSEN

, Stephen BAIN

Graduate School of Medicine,

Kyoto University, Kyoto, Japan;

Billings Clinic Research Center,

Billings, MT, United States of America;

Department of Medicine,

McGill University, Montreal, Canada;

Catalina Research Institute,

Chino, CA, United States of America;

Novo Nordisk A/S, Søborg,

Denmark;

School of Medicine, Swansea University, Wales, United

Kingdom

Semaglutide is a glucagon-like peptide 1 (GLP-1) analog in

development for the treatment of type 2 diabetes (T2D). This

study evaluated the efficacy, safety and tolerability of sub-

cutaneous (s.c.) semaglutide monotherapy versus placebo in

drug-naïve subjects with T2D.

In this phase 3, double-blind study, 388 adults with T2D (HbA1c

–

10%) were randomized to s.c. semaglutide 0.5 mg or 1.0 mg

once weekly or placebo for 30 weeks, including 4

–

8 weeks of

dose escalation. Primary endpoint was change in HbA1c from

baseline to Week 30.

Mean HbA1c (baseline 8.1%) was reduced with semaglutide

0.5 mg and 1.0 mg by 1.5% and 1.6%, respectively, versus <0.1%

in the placebo group (estimated treatment difference versus

placebo [ETD]

−

1.4% and

−

1.5%; p < 0.0001 for both). HbA1c <7%

was achieved by 74% and 72% of 0.5 mg and 1.0 mg semaglu-

tide-treated subjects, versus 25% in the placebo group. The

corresponding proportions of subjects achieving HbA1c

≤

6.5%

were 59%, 60% and 13%. Mean fasting plasma glucose (baseline

9.8 mmol/L) was reduced with semaglutide 0.5 mg and 1.0 mg

by 2.5 mmol/L and 2.3 mmol/L, respectively, versus 0.6 mmol/L

with placebo (ETD

−

2.0 mmol/L and

−

1.8 mmol/L; p < 0.0001

for both).

Mean body weight (BW; baseline 91.9 kg) was significantly

decreased with semaglutide 0.5 mg and 1.0 mg by 3.7 kg and

4.5 kg, respectively, versus 1.0 kg in the placebo group (ETD

−

2.8 kg and

−

3.6 kg; p < 0.0001 for both). Changes in blood

pressure (baseline 128.8/79.3 mmHg) were comparable

between the semaglutide 0.5 mg, 1.0 mg and placebo groups.

Adverse event (AE) and serious AE (SAE) rates were comparable

between groups: 64.1%, 56.2% and 53.5% of patients reported

AEs with semaglutide 0.5 mg, 1.0 mg and placebo, respectively,

and 5.5%, 5.4% and 3.9% reported SAEs. Proportions of patients

discontinuing due to AEs were 6.3%, 5.4% and 2.3% for

semaglutide 0.5 mg, 1.0 mg and placebo. Proportions of sub-

jects reporting gastrointestinal AEs in the 0.5 mg, 1.0 mg and

placebo groups were 20.3%, 23.8% and 7.8% for nausea; 3.9%,

6.9% and 1.6% for vomiting; and 12.5%, 10.8% and 2.3% for

diarrhea. Nausea rate diminished over time.

In conclusion, semaglutide monotherapy, s.c. once-weekly

doses of 0.5 mg and 1.0 mg, significantly improved glycemic

control and reduced BW versus placebo in patients with T2D.

Semaglutide was well tolerated, with a safety profile similar to

that of other GLP-1 receptor agonists.

Poster Presentations / Diabetes Research and Clinical Practice 120S1 (2016) S65

–

S211

S121