Diabetes Research and Clinical Practice - Volume 120 - Supplement 1 - October 2016

74 / 244

underestimate LDL-C especially at high triglyceride levels and
low LDL-C levels. This is especially seen in patients with
diabetes due to abnormalities in lipoprotein composition and
hypertriglyceridemia. A study from our institution derived a
new formula (SMART2D formula) from a cohort of patients
with type 2 diabetes for estimation of LDL-C. The aim of this
study is to validate the use of the SMART2D formula to
estimate LDL-C in patients with and without diabetes.
Methods: We examined 54639 lipid profiles from January 2011
to December 2014 at a Regional General Hospital. LDL-C
calculated using the Friedewald formula (F-LDL-C) and the
SMART2D formula (SMART2D-LDL-C) were compared with
direct LDL-C measurement (M-LDL-C) by an automated assay
(Roche Cobas C501). The agreement in classification into
National Cholesterol Education Program defined LDL levels
was also studied.
Results: In patients with diabetes, the mean difference
between M-LDL-C and F-LDL-C was 0.381 ± 0.253 mmol/L and
between M-LDL-C and SMART2D-LDL-C was −0.148 ± 0.201
mmol/L. Bland Altman plots showed better agreement
between SMART2D-LDL-C and M-LDL-C (B = −0.02, p = 0.160)
than between F-LDL-C and M-LDL-C (B = 0.027, p = 0.00). By
Friedewald formula, 28.7% of patients with M-LDL-C ≥ 2.6
mmol/L were classified as F-LDL-C < 2.6 mmol/L. (Kappa:
0.389) With the SMART2D formula, 5.3% of patients with
M-LDL-C ≥ 2.6 mmol/L were classified as SMART2D-LDL-C
<2.6 mmol/L. (Kappa: 0.828) In patients with TG > 2.2 mmol/L,
45.3% were misclassified with Friedewald formula, while 8.6%
were misclassified with the SMART2D formula.
In patients without diabetes, the mean difference between
M-LDL-C and F-LDL-C was 0.361 ± 0.231 mmol/L and between
M-LDL-C and SMART2D-LDL-C was 0.001 ± 0.182 mmol/L.
Bland Altman plot showed positive bias for the SMART2D
formula (B = −0.004, p = 0.00) and negative bias for the Friede-
wald formula. (B = 0.008, p = 0.00) By Friedewald formula, 33.6%
with M-LDL-C ≥ 2.6 mmol/L were classified as F-LDL-C < 2.6
mmol/L. (Kappa: 0.451) With the SMART2D formula, 6.3% with
M-LDL-C ≥ 2.6 mmol/L were classified as LDL-C < 2.6 mmol/L.
(Kappa: 0.837) In patients with TG > 2.2 mmol/L, 54.2% were
misclassified with Friedewald formula, while 11.4% were
misclassified with the SMART2D formula.
Conclusion: The SMART2D formula, as compared to the
Friedewald formula, provides a more accurate estimate of
LDL-C and reduces misclassification in patients with and
without diabetes.
OL07-5
Elevated hemoglobin A1c levels are associated with increased
arterial stiffness in a Taiwanese population
Chih-Jen CHANG1, Chung-Hao LI2, Yi-Ching YANG1,
Jin-Shang WU 1, Chung-Hung TSAI2, Feng-Hwa LU1.
1 Department of Family Medicine, National Cheng Kung University
Hospital, Tainan, 2Department of Family Medicine, China Medical
University-An Nan Hospital, Taiwan
Introduction: Studies have shown that diabetes mellitus
increased brachial-ankle pulse-wave velocity (baPWV), but
the impact of two pre-diabetes conditions, impaired fasting
glucose and impaired glucose tolerance, remains controver-
sial. Recently, another surrogate for the diagnosis of diabetes
and pre-diabetes has been suggested from an oral glucose
tolerance test to hemoglobin A1c (HbA1c). The aim of this
study was to investigate the impact of different HbA1c status
on baPWV in a relatively healthy Taiwanese population.
Methods: We enrolled 4938 apparently healthy subjects after
excluding those who were under medications for diabetes,
hypertension or hyperlipidemia; or had a history of cardio-
vascular disease; or had diagnosed anemia and peripheral
atherosclerosis with ankle brachial index (ABI) <0.95 in the
health examination center of the National Cheng Kung
University Hospital from Oct. 2006 to Aug. 2009. The baPWV
values to assess arterial stiffness were calculated as the
distance traveled by the pulse wave divided by the time
taken to travel the distance. The participants were classified
into three groups based on 2011 report by an International
Expert Committee, American Diabetes Association: normal
glucose tolerance (NGT) (HbA1C <5.7%, n = 2973), pre-diabetes
(HbA1C 5.7–6.4%, n = 1817) and newly diagnosed diabetes
(NDD) (HbA1C ≥6.5%, n = 148).
Results: The mean values of baPWV were 1265.2 ± 195.0,
1386.7 ± 241.2, 1488.5 ± 278.7 cm/s in NGT, pre-diabetes and
NDD groups, respectively. Both pre-diabetes and NDD groups
had a higher baPWV value as compared with NGT group
(p < 0.001). In multiple linear regression with the reference
group of NGT, both pre-diabetes (β = 13.96, p = 0.002) and NDD
(β = 25.76, p = 0.002) groups had a significantly higher baPWV
values after adjustment for age, sex, body mass index, current
smoking, alcohol consumption, habitual exercise, systolic
blood pressure, cholesterol and high-density lipoprotein
cholesterol.
Conclusions: Diabetic subjects with HbA1c ≥6.5% exhibit a
greater arterial stiffness, even in pre-diabetes with HbA1C of
5.5– 6.4%.
OL07-6
Urine albumin/creatinine ratio is a significant predictor for
incident diabetic peripheral neuropathy in patients with Type
2 Diabetes: 3-year prospective study
Sharon Li Ting PEK3, Na LI3, Lee Ying YEOH 4, Su Chi LIM1–3,
Chee Fang SUM1,2, Subramanian TAVINTHARAN 1–3 . 1Diabetes
Centre, Khoo Teck Puat Hospital, 2Division of Endocrinology, Khoo
Teck Puat Hospital,3Clinical Research Unit, Khoo Teck Puat Hospital,
4Division of Nephrology, Khoo Teck Puat Hospital, Singapore
Introduction: Diabetic peripheral neuropathy (DPN) is a
common complication of Type 2 Diabetes (T2D). Apart from
hyperglycemia, its pathogenesis is poorly understood. DPN
can develop despite intensive glucose control, suggesting that
other risk factors are involved in the pathophysiology of DPN.
In this study, we aim to determine baseline predictors of
incident DPN.
Methods: We consecutively enrolled patients (n = 2058) with
T2D (21–90 years old), seen our institution’s Diabetes Centre
and a primary care polyclinic in Singapore (August 2011–March
2014). From 2014, 684 of the participants made a repeat visit
3-year from first visit. The prospective data collection is still
on-going. Anthropometric data, fasting blood, urine were
collected for biochemistry and urine albumin/creatinine
measurements (uACR). Systolic and diastolic blood pressure
(SBP and DBP) was taken fromparticipants using an automated
blood pressure monitor. Carotid-femoral Pulse wave velocity
(PWV) was determined using SphygmoCor equipment and
software. Neuropathy was considered present if an abnormal
finding in monofilament (<8 of 10 points) or neurothesiometer
testing ≥25 volts on at least one foot.
Results: Baseline characteristics of 684 patients are: Age:
(57.2 ± 10.4) years, 50.1% males, ethnicity: 53.8% Chinese,
20.9% Malay, 21.5% Indian and 3.8% of others, duration of
T2D: (11.1 ± 8.6) years and HbA1c (7.71 ± 1.26) mmol/L. In these
684 patients, 599 had no DPN at baseline. On a 3-year follow-
up, 48 (8%) of them developed DPN. Patients with incident DPN
(n = 48) versus controls who did not developDPN (n = 551), were
older [(60.4 ± 8.0) vs (56.9 ± 10.6)]years (p = 0.026), had longer
duration of diabetes [(14.6 ± 10.2) vs (10.4 ± 8.2)] years (p =
0.007), increased SBP [(148.4 ± 16.5) vs (136.6 ± 17.6)] mmHg
(p < 0.0001) and DBP [(81.1 ± 9.1) vs (77.7 ± 9.4)] mmHg (p =
0.017), increased PWV [(10.8 ± 2.5) vs (9.6 ± 2.6)] m/s (p = 0.002),
lower eGFR [(75.7 ± 34.8) vs (90.5 ± 29.7)] (p = 0.001), worse uACR
[18.0(6.0–63.8) vs 51.0(20.0–334.0)] mg/g (p < 0.0001) at baseline.
There were no significant difference for patients, with incident
DPN vs those without, in their baseline HbA1c [(7.81 ± 1.35) vs
(7.65 ± 1.25)] mmol/L and fasting glucose [(7.88 ± 2.43) vs
Oral Presentations / Diabetes Research and Clinical Practice 120S1 (2016) S40–S64 S56