the association between serum uric acid and residual-cell
TRANSCRIPT
Research ArticleThe Association between Serum Uric Acid and Residual š½-CellFunction in Type 2 Diabetes
Wei Tang,1 Qi Fu,2 Qingqing Zhang,2 Min Sun,2 Yuan Gao,1
Xuan Liu,2 Li Qian,2 Shan Shan,2 and Tao Yang2
1 Department of Endocrinology and Metabolism, The Affiliated Jiangyin Hospital of Southeast University Medical College,Jiangyin, Jiangsu 214400, China
2Department of Endocrinology, The First Affiliated Hospital of Nanjing Medical University, Nanjing, Jiangsu 210029, China
Correspondence should be addressed to Wei Tang; [email protected] and Tao Yang; [email protected]
Received 16 November 2013; Revised 14 April 2014; Accepted 7 May 2014; Published 26 May 2014
Academic Editor: Li Yanbing
Copyright Ā© 2014 Wei Tang et al.This is an open access article distributed under the Creative CommonsAttribution License, whichpermits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
The aim of this study was to investigate the relationship of serum uric acid (sUA) with residual š½-cell function in type 2 diabetes.Oral glucose tolerance tests (OGTT) were performed on 1021 type 2 diabetes patients. The ratio of area under curve of insulin toglucose during 0 to 30min and 0 to 120min of the OGTT was calculated as indices of insulin secretion function. The productsof insulin secretion indices multiplied by Matsuda insulin sensitivity index were used as disposition indices. After correlation andmultiple linear regression analysis, sUA was significantly associated with insulin secretion and disposition indices in male, female,and total groups adjusted for confounding factors (includingmetabolic indicators like sex, age, course of the disease, blood glucose,blood pressure, serum lipids, and so on). Superficially steeper time-dependent decline of insulin secretion function was found inpatients with sUA above the median than those below it. In conclusion, our results suggest an independent positive associationbetween sUA and residual š½-cell function in type 2 diabetes. Patients with higher sUA have greater insulin secretion ability thanthose with lower sUA at the early stage of disease, but their residual š½-cell function seems to decay more rapidly.
1. Introduction
Lots of studies found increased serum uric acid (sUA) levelsin subjects with metabolic syndrome (MetS) or cardiovascu-lar disease, and sUA is associated with several componentsof MetS, including dyslipidemia, hypertension, impairedglucose metabolism, and obesity [1ā3]. Some recent studieshave already highlighted the connection between sUA andglucose homeostasis. For example, the Rotterdam Studyshowed that the subjects with higher levels of sUA were athigher risk of type 2 diabetes [4], and a modest positiveassociation between plasma uric acid concentration and theincidence of type 2 diabetes in Chinese individuals wassuggested by Chien et al. [5]. As is known, both insulinresistance and š½-cell dysfunction play determinate roles inthe pathogenesis of type 2 diabetes [6]. The correlationbetween sUA and insulin resistance has been verified byseveral researches years ago [7, 8], and insulin resistanceis thought to be the principal underlying pathophysiologic
abnormality connecting hyperuricemia and components ofMetS. However, most previous studies were performed onnondiabetic subjects. Furthermore, few studies focused onthe relationship between sUA level and islet š½-cell dysfunc-tion, and whether there is a corresponding change of sUAlevel with š½-cell function deterioration in type 2 diabetes isunknown. In this cross-sectional study, we investigated theassociation between sUAandš½-cell function aswell as insulinsensitivity in patients with type 2 diabetes, and we furtherelucidate the time-dependent changes of insulin secretionability in different gender and uric acid level groups.
2. Materials and Methods
2.1. Study Subjects. One thousand and twenty-one patientswith type 2 diabetes who received treatment in the First Affil-iated Hospital of Nanjing Medical University between 2008and 2011 were enrolled in this study. Type 2 diabetes mellitus
Hindawi Publishing CorporationJournal of Diabetes ResearchVolume 2014, Article ID 709691, 9 pageshttp://dx.doi.org/10.1155/2014/709691
2 Journal of Diabetes Research
Table1:Ba
selin
echaracteristicso
fsub
jectsg
roup
edby
gend
erandserum
uricacid
levels(m
eanĀ±SD
ornu
mber).
Female
Male
Total
LUA(š=209)
HUA
(š=210)
LUA(š=301)
HUA
(š=301)
LUA(š=510)
HUA
(š=511)
sUA(šmol/L)
207.61Ā±38.31
339.84Ā±71.25#
250.19Ā±46.47
383.06Ā±61.04#
232.74Ā±48.08
365.29Ā±68.73#
Age
(yr)
57.83Ā±11.11
61.40Ā±11.42ā
55.65Ā±11.63
54.23Ā±13.52
56.54Ā±11.45
57.18Ā±13.16
Yearsfrom
diagno
sis6.19Ā±6.36
6.24Ā±5.72
5.82Ā±6.22
5.16Ā±5.72
5.97Ā±6.27
5.60Ā±5.74
Hypertension(yes/no)
85/124
146/64
#115/186
149/152ā
200/310
295/216#
FDM
(yes/no)
78/131
80/130
123/178
113/188
201/309
213/308
SBP(m
mHg)
133.04Ā±16.71
135.33Ā±16.52
131.33Ā±14.95
133.00Ā±16.15
132.03Ā±15.70
133.96Ā±16.33
DBP
(mmHg)
81.46Ā±9.35
81.85Ā±10.29
82.56Ā±9.83
83.94Ā±9.74
82.11Ā±9.65
83.08Ā±10.01
BMI(kg/m
2 )24.39Ā±3.67
25.87Ā±3.62
#24.24Ā±3.26
25.92Ā±3.22
#24.29Ā±3.43
25.90Ā±3.38
#
HbA
1c(%
)9.51Ā±2.51
8.85Ā±2.39ā
9.83Ā±2.55
9.15Ā±2.39ā
9.70Ā±2.53
9.02Ā±2.39
#
TC(m
mol/L)
4.99Ā±1.18
5.09Ā±1.15
4.70Ā±1.13
4.74Ā±1.10
4.81Ā±1.15
4.88Ā±1.13
TG(m
mol/L)
1.57Ā±1.17
1.90Ā±1.30
#1.57Ā±1.39
2.27Ā±2.21
#1.56Ā±1.30
2.11Ā±1.89
#
HDL(m
mol/L)
1.25Ā±0.32
1.17Ā±0.28ā
1.15Ā±0.32
1.01Ā±0.23
#1.19Ā±0.32
1.07Ā±0.26
#
LDL(m
mol/L)
3.12Ā±0.87
3.19Ā±0.84
3.00Ā±0.79
3.03Ā±0.74
3.05Ā±0.82
3.09Ā±0.78
ALT
(IU/L)
25.17Ā±18.35
28.52Ā±25.71
25.65Ā±18.95
33.65Ā±25.17#
25.45Ā±18.69
31.54Ā±25.49#
INSR
301.31Ā±1.01
1.78Ā±1.13
#0.99Ā±0.76
1.56Ā±1.12
#1.12Ā±0.88
1.65Ā±1.12
#
INSR
120
1.59Ā±1.43
2.29Ā±1.76
#1.18Ā±1.08
1.94Ā±1.59
#1.34Ā±1.25
2.08Ā±1.66
#
DI30
7.10Ā±3.89
7.35Ā±3.59
6.49Ā±3.20
7.47Ā±3.65
#6.74Ā±3.50
7.41Ā±3.62
#
DI120
8.52Ā±5.78
9.10Ā±4.94
7.61Ā±4.35
9.23Ā±5.46
#7.97Ā±5.00
9.17Ā±5.24
#
HOMA-
IR2.46Ā±1.73
3.04Ā±1.86
#2.02Ā±1.48
2.71Ā±1.89
#2.20Ā±1.59
2.84Ā±1.88
#
Matsuda
ISI
7.27Ā±4.71
5.17Ā±3.11
#8.99Ā±5.48
6.36Ā±3.93
#8.28Ā±5.24
5.87Ā±3.65
#
Abno
rmallydistr
ibuted
continuo
usvaria
bles,including
HbA
1c,T
G,A
LT,INS0,INS30,IN
S120,INSR
30,INSR
120,DI30,DI120,H
OMA-
IR,and
Matsuda
ISI,werelog-transfo
rmed
fora
nalysis;n
ontransfo
rmed
values
wered
isplay
edfore
aseo
finterpretation.
Differencesb
etweentheL
UAandHUA
grou
pweree
stimated
usingun
paire
dStud
entāsš”-test(orc
hi-squ
aretestfor
classified
data).ā š<0.01;#š<0.001.
Journal of Diabetes Research 3
was diagnosed according to the criteria of the AmericanDiabetes Association [9].Themaximum duration of diabeteswas 35 years.The average age of all patients was 56.86 Ā± 12.34(mean Ā± SD) years old. Meanwhile, body mass index (BMI)and sUA were 25.10 Ā± 3.50 kg/m2 and 299.08 Ā± 88.96 šmol/L,respectively. Patients with severe pancreatic disease, liverdisease, and renal disease and those who suffered recentdiabetic ketoacidosis and hyperosmotic nonketotic diabeticcoma were excluded. Verbal informed consent was obtainedfrom all participants. The study was approved by the ethicscommittee of the First Affiliated Hospital of Nanjing MedicalUniversity.
2.2.Measurements. In all subjects, the height, weight, systolicblood pressure (SBP), and diastolic blood pressure (DBP)weremeasured and recorded. History of hypertension, familyhistory of diabetes (FHD), and the years from diagnosisof type 2 diabetes were inquired. If one of parents, sib-lings, or grandparents had been diagnosed with diabetes,patient was defined as having FHD. All patients stoppedusing antidiabetic medicine at least one day before theblood samples are taken. After 10ā12 hours overnight fast-ing, venous blood samples were collected to measure uricacid, glycated haemoglobin (HbA1c), triglyceride (TG), totalcholesterol (TC), high-density lipoprotein cholesterol (HDL),low-density lipoprotein cholesterol (LDL), alanine amino-transferase (ALT), fasting plasma glucose (G0), and fastingserum insulin (I0). Then the 75 g oral glucose tolerancetest (OGTT) and insulin secretion test were performed, andvenous blood samples were obtained at 30 and 120 minutesafter glucose load for measuring the plasma glucose (G30,G120) and serum insulin (I30, I120).
Plasma glucose concentrations were measured using thehexokinasemethod (OLYMPUSAU5400). HbA1c and seruminsulin were measured by high performance liquid chro-matography (Bio-Rad D10) and radioimmunoassay (Iodine[125I] Insulin Radioimmunoassay Kit, Beijing North Instituteof Biological Technology), respectively. ALT and serum lipidprofiles, including TG, HDL, and LDL, were determined withan automatic biochemical analyzer (HITACHI 7020).
2.3. Calculations. BMI was calculated through dividingweight (kg) by square of height (m). To evaluate the insulinsecretion, InsAUC30/GluAUC30 (INSR30) was calculatedas a surrogate index for the early phase insulin secre-tion and InsAUC120/GluAUC120 (INSR120) as a surrogateindex for total insulin secretion, where InsAUC30 andGluAUC30 are the area under insulin (mIU/L) and glucose(mmol/L) curves during 0 to 30min of the OGTT andInsAUC120 and GluAUC120 are the area under insulinand glucose curves during 0 to 120min, respectively [10].Matsuda insulin sensitivity index (Matsuda ISI, calculated as10000/ā(G0 Ć I0) Ć (G Ć I), where G and I are the averagelevels of plasma glucose (mg/dL) and insulin (mIU/L) duringOGTT) was chosen to evaluate insulin sensitivity [11]. Thehomeostasis model assessment of insulin resistance (HOMA-IR) indexwas calculated as follows: insulin (mIU/L)Ć glucose(mmol/L)/22.5 [12]. Glucose disposition indices (disposition
Table 2: Correlations between serum uric acid level and š½-cellfunction as well as insulin sensitivity.
Total Female Maleš š š
INSR30 0.301# 0.290# 0.395#
INSR120 0.296# 0.275# 0.393#
DI30 0.150# 0.100ā 0.206#
DI120 0.166# 0.115ā 0.226#
HOMA-IR 0.192# 0.216# 0.241#
Matsuda ISI ā0.252# ā0.261# ā0.336#
All abnormally distributed continuous variables were log-transformed. āš <0.05; #š < 0.001.
index 30 = DI30 = Matsuda ISI Ć INSR30, disposition index120 = DI120 = Matsuda ISI Ć INSR120) [10, 13] were used toassess š½-cell function, combining both insulin secretion andinsulin sensitivity.
2.4. Statistical Analysis. Studentās t-test and the chi-squaretest were used to analyze group differences. The correla-tivity between sUA and š½-cell function was analyzed byPearsonās correlation. The multiple linear regression analysiswas applied to test the associations between š½-cell functionand sUA after adjustment for several covariates. Abnormallydistributed continuous variables, including HbA1c, TG, andALT, as well as all the indices for insulin sensitivity and š½-cell function, were log-transformed to yield an approximatelynormal distribution before statistical analysis. š value < 0.05(two-tailed) was considered statistically significant. All statis-tical analyses were conducted with the Statistical Package forSocial Science for Windows (SPSS, version 13.0).
3. Results
The characteristic of the study patients was shown in Table 1.The patients were divided into two groups according to themedian sUA levels of females and males, respectively (LUA:low serum uric acid, which was under the median sUA level;HUA: high serum uric acid, which was above the mediansUA level). Patients with HUA had higher levels of BMI, TG,and ALT and greater ratio of hypertension in both genders.In contrast, the HbA1c and HDL were lower in HUA groupsthan in LUA ones. In male, female, and total groups, HUApatients had greater HOMA-IR and smallerMatsuda ISI thanLUA subjects. Consistent with insulin resistance, patientswith HUA had greater insulin secretion indices, that is,INSR30 and INSR120, than LUA in all groups. Interestingly,we found that in men and total groups, patients with HUAhad greater disposition indices (both DI30 and DI120) ratherthan in women group. Statistically significant correlationswere found between sUA level and all the indices of š½-cellfunction and insulin sensitivity either totally or after beingstratified by gender (shown in Table 2).
To exclude confounding factors which may influence š½-cell function, multiple linear regression was carried out usingall clinical parameters which were significantly correlated
4 Journal of Diabetes Research
Table 3: Multiple linear regression for š½-cell function in total patients.
INSR30 INSR120 DI30 DI120š½ S-š½ š½ S-š½ š½ S-š½ š½ S-š½
Sex (0 = female, 1 = male) ā0.201 ā0.136# ā0.229 ā0.139# ā ā ā āHypertension (yes = 1, no = 0) ā0.026 ā0.018 ā0.018 ā0.011 ā ā ā āYears from diagnosis ā0.020 ā0.165# ā0.029 ā0.218# ā0.016 ā0.217# ā0.026 ā0.287#
BMI (kg2/cm) 0.021 0.005# 0.018 0.076ā ā ā ā āSBP (mmHg) ā ā ā ā 0.002 0.055 0.001 0.021DBP (mmHg) ā ā ā ā ā0.005 ā0.099ā ā0.006 ā0.111#
HbA1c (%) ā0.116 ā0.400# ā0.156 ā0.479# ā0.075 ā0.414# ā0.114 ā0.527#
sUA (šmol/L) 0.001 0.147# 0.001 0.146# 0.001 0.154# 0.001 0.156#
TG (mmol/L) ā0.006 ā0.005 0.015 0.012 ā ā ā āTC (mmol/L) ā ā ā ā ā0.011 ā0.029 ā0.019 ā0.040HDL (mmol/L) ā0.077 ā0.032 ā0.098 ā0.036 ā ā ā āLDL (mmol/L) ā ā ā ā ā0.007 ā0.012 0.009 0.014ALT (IU/L) 0.091 0.076ā 0.103 0.077ā ā ā ā āHOMA-IR 0.497 0.455# 0.426 0.349# ā0.222 ā0.326# ā0.279 ā0.343#
Ī: partial regression coefficient; S-š½: standard partial regression coefficient. All abnormally distributed continuous variables were log-transformed. Onlyparameters which were independently associated with š½-cell function indices after multiple stepwise regression are shown. ā š < 0.01; #š < 0.001.
with š½-cell function indices (data of correlation was notshown) as independent variables and indices of š½-cell func-tion as dependent variables. The results of multiple linearregression analysis are shown in Table 3. The sUA was inde-pendently associated with INSR30 and INSR120 (š value ofboth partial regression coefficients was less than 0.001) afteradjustment for sex, hypertension, years from diagnosis, BMI,HbA1c, TG, HDL, ALT, and HOMA-IR. In the regressionmodel of DI30 and DI120, the relationship between sUAand disposition indices was statistically significant (š valueof both partial regression coefficients was less than 0.001)after adjustment for years from diagnosis, SBP, DBP, HbA1c,TC, LDL, and HOMA-IR. After grouping patients by sex,the regression models of š½-cell function were shown inTable 4. Serum uric acid remained significantly associatedwith insulin secretion indices (INSR30 and INSR120) anddisposition indices (DI30 and DI120) after adjusting forpotential confounding factors in both female and malegroups (all š value of partial regression coefficients was lessthan 0.05).
We also investigated the differences of š½-cell functionchanges along with disease duration between low and highsUA levels. In the scattered plots displayed in Figures 1 and 2,in both females and males, INSR30 and INSR120 decreasedalong with disease duration in both HUA and LUA group (allregression coefficients were negative, š < 0.05). Interestingly,although INSR30 and INSR120 were higher in HUA groupat the early stage of diabetes, they superficially decreasedmore rapidly along with disease duration than in LUA groupand finally dropped to almost the same level as the LUAgroup. However, the difference of slop of the lines was notstatistically significant after adjusting for confounding factorsincluding hypertension, BMI, HbA1c, TG, HDL, ALT, andHOMA-IR. For disposition indices, in HUA group, DI30 andDI120 significantly dropped as disease duration increasedin both females and males. However, in LUA group, no
significant regression was found between disposition indicesand disease duration in both genders, except for DI120 infemales.
4. Discussion
The failure of pancreatic š½-cell function plays an importantrole in the pathogenesis of type 2 diabetes. Previous studieshave shown that impaired insulin secretion is the key inthe conversion from normal glucose tolerance (NGT) toimpaired glucose tolerance (IGT) and diabetes [14, 15], andthe deterioration of š½-cell function does not stop afterdiagnosis. The United Kingdom Prospective Diabetes Study(UKPDS) showed that š½-cell function, assessed by home-ostasis model assessment (HOMA), decreased approximatelyby 25% in the first 5 years of diabetes [16]. Several factors,including hyperglycemia, dyslipidemia, cytokines secretedby adipocytes, and immune response, have been proposedas reasons for pancreatic š½-cell function deterioration [17ā19]. In addition, our study found the close relationshipbetween sUA and insulin secretion ability as well as glucosedisposition indices, which was rarely concerned in previousstudies.
Concerning the relationship between sUA level and twocritical sides in the pathogenesis of type 2 diabetes, that is,insulin resistance and š½-cell dysfunction, the interrelation-ship between sUA and insulin resistance was revealed byseveral studies years ago [7, 8]. Our present study verifiedthis relationship again by calculating the insulin sensitivity byeither Matsuda ISI or HOMA-IR. Elevated sUA level usuallyaccompanies insulin resistance. Higher insulin levels canreduce renal excretion of urate and enhance renal urate reab-sorption with increased renal tubular reabsorption of sodium[7, 20, 21]. In addition, increased purine biosynthesis andturnover, with its attendant increase in sUA, link high sUA
Journal of Diabetes Research 5
Table4:Multip
lelin
earregressionforš½
-cellfun
ctionaft
ergrou
ping
bygend
er.
INSR
30IN
SR120
DI30
DI120
š½S-š½
š½S-š½
š½S-š½
š½S-š½
Female
Hypertension(yes
=1,no
=0)
ā0.016
ā0.012
ā0.023
ā0.015
āā
āā
Yearsfrom
diagno
sisā0.024
ā0.216#
ā0.034
ā0.268#
ā0.020
ā0.256#
ā0.030
ā0.331#
BMI(kg
2 /cm
)0.015
0.082ā
0.010
0.050
āā
āā
SBP(m
mHg)
āā
āā
0.001
0.051
0.001
0.008
DBP
(mmHg)
āā
āā
ā0.00
6ā0.125ā
ā0.00
6ā0.111ā
HbA
1c(%
)ā0.114
ā0.40
9#ā0.155
ā0.490#
ā0.069
ā0.366#
ā0.109
ā0.491#
sUA(šmol/L)
0.001
0.115ā
0.001
0.107ā
0.001
0.143ā
0.001
0.143#
TG(m
mol/L)
0.021
0.017
0.04
60.032
āā
āā
TC(m
mol/L)
āā
āā
ā0.052
ā0.130
ā0.074
ā0.159
HDL(m
mol/L)
ā0.065
ā0.029
ā0.131
ā0.052
āā
āā
LDL(m
mol/L)
āā
āā
0.039
0.070
0.063
0.098
ALT
(IU/L)
0.106
0.096ā
0.117
0.094ā
āā
āā
HOMA-
IR0.456
0.435#
0.412
0.345#
ā0.275
ā0.384#
ā0.309
ā0.367#
Male Hypertension(yes
=1,no
=0)
ā0.024
ā0.016
ā0.00
6ā0.00
4ā
āā
āYearsfrom
diagno
sisā0.017
ā0.134#
ā0.026
ā0.186#
ā0.014
ā0.187#
ā0.023
ā0.258#
BMI(kg
2 /cm
)0.026
0.116
#0.024
0.100ā
āā
āā
SBP(m
mHg)
āā
āā
0.001
0.034
0.001
0.021
DBP
(mmHg)
āā
āā
ā0.003
ā0.056
ā0.005
ā0.086ā
HbA
1c(%
)ā0.116
ā0.395#
ā0.154
ā0.474#
ā0.075
ā0.423#
ā0.112
ā0.526#
sUA(šmol/L)
0.001
0.163#
0.002
0.169#
0.001
0.201#
0.001
0.212#
TG(m
mol/L)
ā0.026
ā0.023
0.001
0.001
āā
āā
TC(m
mol/L)
āā
āā
0.00
40.033
0.002
0.003
HDL(m
mol/L)
ā0.088
ā0.034
ā0.06
4ā0.022
āā
āā
LDL(m
mol/L)
āā
āā
ā0.036
ā0.063
ā0.021
ā0.030
ALT
(IU/L)
0.075
0.06
0ā0.083
0.06
0āā
āā
āHOMA-
IR0.522
0.469#
0.430
0.350#
ā0.202
ā0.304#
ā0.283
ā0.353#
Ī:partia
lregressioncoeffi
cient;S-š½:stand
ardpartialregressioncoeffi
cient.Allabno
rmallydistr
ibuted
continuo
usvaria
bleswerelog-tr
ansfo
rmed.O
nlyp
aram
etersw
hich
wereind
ependentlyassociated
withš½-cell
functio
nindicesa
fterm
ultip
leste
pwise
regressio
nares
hown.āš<0.05;ā š<0.01;#š<0.001.
6 Journal of Diabetes Research
35302520151050
3
2
1
0
ā1
ā2
ā3
InIN
SR30
(IU
/mol
)
Years from diagnosisLUA HUAy = ā0.021x + 0.166 y = ā0.031x + 0.581
(a)
35302520151050
3
2
1
0
ā1
ā2
ā3
Years from diagnosis
InIN
SR120
(IU
/mol
)LUAy = ā0.028x + 0.341
HUAy = ā0.040x + 0.820
(b)
3.5
3.0
2.5
2.0
1.5
1.0
0.0
0.5
35302520151050
Years from diagnosisLUAHUA
y = ā0.020x + 2.012
InD
I30(L
2/(
mgā
mm
ol))
(c)
4.0
3.5
3.0
2.5
2.0
1.5
1.0
0.5
35302520151050
Years from diagnosis
LUAy = ā0.015x + 2.065
HUAy = ā0.029x + 2.251
InD
I120(L
2/(
mgā
mm
ol))
(d)
Figure 1: Scattered plots of simple linear regression between š½-cell function and years from diagnosis in females. LUA, sUA levels undermedian of females; HUA, sUA levels above median of females. All š½-cell function indices (INSR30, INSR30, DI30, and DI120) were log-transformed. Only significant regression lines and formulas are shown. All regression coefficients were negative. The decreases of INSR30,INSR120, and DI120 per year in the HUA group were greater than that in LUA group.
to insulin resistance and/or hyperinsulinaemia by increasedactivity of the hexose monophosphate shunt [22]. On theother hand, sUA not only may be a consequence of insulinresistance but also may actually promote or worsen insulinresistance. Specifically, a recent study showed that sUA plays
an important role in the pathogenesis of MetS, possibly dueto its ability to inhibit endothelial function. In detail, sUAhas been shown to inhibit nitric oxide (NO) bioavailabilityand reduce NO concentration which is required in insulinstimulated glucose uptake [23, 24]. Consequently, higher sUA
Journal of Diabetes Research 7
302010 400
3
2
1
0
ā1
ā2
ā3
InIN
SR30
(IU
/mol
)
Years from diagnosisLUAy = ā0.014x ā 0.177
HUAy = ā0.022x + 0.335
(a)
302010 400
Years from diagnosis
3
2
1
0
ā1
ā2
ā3
InIN
SR120
(IU
/mol
)HUALUA
y = ā0.016x ā 0.050 y = ā0.033x + 0.558
(b)
302010 400
Years from diagnosis
3.5
3.0
2.5
2.0
1.5
1.0
0.5
0.0
LUAHUA
y = ā0.011x + 1.969
InD
I30(L
2/(
mgā
mm
ol))
(c)
4.0
3.5
3.0
2.5
2.0
1.5
1.0
0.5
Years from diagnosis
LUAHUA
302010 400
y = ā0.022x + 2.191
InD
I120(L
2/(
mgā
mm
ol))
(d)
Figure 2: Scattered plots of simple linear regression betweenš½-cell function and years fromdiagnosis inmales. LUA, sUA levels undermedianof men; HUA, sUA levels above median of males. All š½-cell function indices (INSR30, INSR30, DI30, and DI120) were log-transformed. Onlysignificant regression lines and formulas are shown. All regression coefficients were negative. The decreases of INSR30 and INSR120 per yearin the HUA group were greater than that in LUA group.
always keeps with more severe insulin resistance, that is,greater insulin demand of our organs.
Few previous researches investigated the interactionbetween š½-cell function and sUA. In present study, wereported the correlation between sUA level and indices
reflecting islet insulin secretion ability. Subjects with higherlevels of sUA had higher insulin secretion, including theearly phase (INSR30) and total (INSR120) insulin secretion,and after adjustment for variables associated with insulinresistance including BMI, TG, ALT, and HOMA-IR, sUA
8 Journal of Diabetes Research
is still independently associated with insulin secretion. Thereason why patients with type 2 diabetes of HUA groupcan secrete more insulin at the early stage was temporarilyunknown. Insulin resistance may be one of the supposi-tional reasons, and the increased insulin secretion may beconsidered as a compensatory response to overcome theinsulin resistance. Although subjects with higher sUA secretemore insulin, it does not mean that high sUA is beneficialto š½-cell function. The insulin secretion ability in thoseseems to drop more rapidly than those having lower sUA astype 2 diabetes duration extends. At last, when the diseaseduration is long enough, the difference of both INSR30and INSR120 between HUA and LUA groups diminishes.A recent study provided evidence that sUA has a directnegative effect on š½-cell function, which could cause š½-celldeath and dysfunction by activation of the NF-š B and iNOS-NO signal axis. This may partly explain the reason whyinsulin secretion ability seems to drop more rapidly in HUAgroup [25]. Nevertheless, the profound physiology changesconnecting uric acid metabolism and insulin secretion is stillworth further investigation.
Disposition indices which reflect the real glucosemetabolism give attention to both insulin sensitivity andinsulin secretion after glucose load. In our research, the sUAalways kept positive association with disposition indicesin total patients and after having been divided into gendergroups. This finding is consistent with the result also got inour study that those subjects with higher sUA got statisticallylower HbA1c level. All these suggest that high sUA levelis associated with better glucose utilization. A mechanismunderlying the relationship between glucose utilization andsUA levels may be due to the uricosuric effect of glycosuria,which means hyperglycemia facilitates uric acid excretionwhen the blood glucose level is above 10mmol/L [26]. Byglycosuria, the sUA concentration of patients with highblood glucose levels may be low; however high bloodglucose is harmful to islet š½-cells. Furthermore, some studiesdemonstrated high sUA levels are associated with increasedgeneration of free radicals and oxidative stress [27], whichhas various adverse effects on š½-cell function [28, 29]. Thiscould be one possible reason for the faster decline of insulinsecretion function in high sUA patients. However, otherstudies have suggested that sUA is an effective antioxidant[30, 31] and elevated sUA levels may reflect a compensatorymechanism contributing to the increased oxidative stressassociated with the MetS. Collectively, the exact role of sUAin oxidation is still controversial, and further research isrequired.
We are very cautious about making conclusion in theclose relation between sUA and š½-cell insulin secretionfunction as well as glucose disposition because of two points:firstly, this is just a cross-sectional study rather than longitu-dinal study; secondly, subjects in our study received a widediversity of hypoglycemic therapy which might influence thenature change of the islet function.
However, from this study, an independent positive asso-ciation between sUA and š½-cell function is confirmed, sug-gesting a potential close relation and interaction between uricacid and insulin secretion ability. Type 2 diabetic patients
with higher sUA level have better insulin secretion but theirresidual š½-cell function seems to decay more quickly.
Conflict of Interests
The authors declare that there is no conflict of interestsregarding the publication of this paper.
Acknowledgment
The authors would like to thank Chong Shen, Departmentof Epidemiology and Biostatistics, School of Public Health,Nanjing Medical University, Nanjing, China, for statisticalassistance.
References
[1] N. Ishizaka, Y. Ishizaka, E.-I. Toda, R. Nagai, andM. Yamakado,āAssociation between serumuric acid,metabolic syndrome, andcarotid atherosclerosis in Japanese individuals,āArteriosclerosis,Thrombosis, and Vascular Biology, vol. 25, no. 5, pp. 1038ā1044,2005.
[2] I.-C. Hwang, S.-Y. Suh, A.-R. Suh, and H.-Y. Ahn, āTherelationship between normal serum uric acid and nonalcoholicfatty liver disease,ā Journal of Korean Medical Science, vol. 26,no. 3, pp. 386ā391, 2011.
[3] E. S. Ford, C. Li, S. Cook, and H. K. Choi, āSerum concen-trations of uric acid and the metabolic syndrome among USchildren and adolescents,ā Circulation, vol. 115, no. 19, pp. 2526ā2532, 2007.
[4] A. Dehghan, M. van Hoek, E. J. G. Sijbrands, A. Hofman, and J.C.M.Witteman, āHigh serum uric acid as a novel risk factor fortype 2 diabetes,ā Diabetes Care, vol. 31, no. 2, pp. 361ā362, 2008.
[5] K.-L. Chien,M.-F. Chen, H.-C. Hsu et al., āPlasma uric acid andthe risk of type 2 diabetes in a Chinese community,ā ClinicalChemistry, vol. 54, no. 2, pp. 310ā316, 2008.
[6] S. E. Kahn, āThe relative contributions of insulin resistanceand beta-cell dysfunction to the pathophysiology of Type 2diabetes,ā Diabetologia, vol. 46, no. 1, pp. 3ā19, 2003.
[7] F. Facchini, Y.-D. I. Chen, C. B. Hollenbeck, and G. M. Reaven,āRelationship between resistance to insulin-mediated glucoseuptake, urinary uric acid clearance, and plasma uric acidconcentration,ā Journal of the American Medical Association,vol. 266, no. 21, pp. 3008ā3011, 1991.
[8] S. G. Tsouli, E. N. Liberopoulos, D. P.Mikhailidis, V. G. Athyros,and M. S. Elisaf, āElevated serum uric acid levels in metabolicsyndrome: an active component or an innocent bystander?āMetabolism: Clinical and Experimental, vol. 55, no. 10, pp. 1293ā1301, 2006.
[9] American Diabetes Association, āDiagnosis and classificationof diabetes mellitus,ā Diabetes Care, vol. 30, no. 1, supplement,pp. S42āS47, 2007.
[10] A. Staacakova, M. Javorsky, T. Kuulasmaa, S. M. Haffner, J.Kuusisto, and M. Laakso, āChanges in insulin sensitivity andinsulin release in relation to glycemia and glucose tolerance in6,414 finnish men,ā Diabetes, vol. 58, no. 5, pp. 1212ā1221, 2009.
[11] M. Matsuda and R. A. DeFronzo, āInsulin sensitivity indicesobtained from oral glucose tolerance testing: comparison withthe euglycemic insulin clamp,ā Diabetes Care, vol. 22, no. 9, pp.1462ā1470, 1999.
Journal of Diabetes Research 9
[12] D. R. Matthews, J. P. Hosker, A. S. Rudenski et al., āHomeostasismodel assessment: insulin resistance and š½-cell function fromfasting plasma glucose and insulin concentrations in man,āDiabetologia, vol. 28, no. 7, pp. 412ā419, 1985.
[13] R. Retnakaran, S. Shen, A. J. Hanley, V. Vuksan, J. K. Hamilton,and B. Zinman, āHyperbolic relationship between insulin secre-tion and sensitivity on oral glucose tolerance test,ā Obesity, vol.16, no. 8, pp. 1901ā1907, 2008.
[14] C. Weyer, C. Bogardus, D. M. Mott, and R. E. Pratley, āThenatural history of insulin secretory dysfunction and insulinresistance in the pathogenesis of type 2 diabetes mellitus,āJournal of Clinical Investigation, vol. 104, no. 6, pp. 787ā794,1999.
[15] S. E. Kahn, āClinical, review 135: the importance of š½-cell failurein the development and progression of type 2 diabetes,ā Journalof Clinical Endocrinology and Metabolism, vol. 86, no. 9, pp.4047ā4058, 2001.
[16] United Kingdom Prospective Diabetes Study Group, āU.K.Prospective Diabetes Study 16. Overview of 6 years therapy oftype II diabetes: a progressive disease,ā Diabetes, vol. 44, no. 11,pp. 1249ā1258, 1995.
[17] S. del Prato, āRole of glucotoxicity and lipotoxicity in thepathophysiology of Type 2 diabetes mellitus and emergingtreatment strategies,āDiabeticMedicine, vol. 26, no. 12, pp. 1185ā1192, 2009.
[18] M. Y. Donath, D. J. Gross, E. Cerasi, and N. Kaiser,āHyperglycemia-induced š½-cell apoptosis in pancreatic islets ofPsammomys obesus during development of diabetes,āDiabetes,vol. 48, no. 4, pp. 738ā744, 1999.
[19] K. Maedler, G. A. Spinas, D. Dyntar, W. Moritz, N. Kaiser, andM. Y. Donath, āDistinct effects of saturated and monounsatu-rated fatty acids on š½-cell turnover and function,āDiabetes, vol.50, no. 1, pp. 69ā76, 2001.
[20] H. K. Choi, D. B. Mount, and A. M. Reginato, āPathogenesis ofgout,ā Annals of Internal Medicine, vol. 143, no. 7, pp. 499ā516,2005.
[21] F. P. Cappuccio, P. Strazzullo, E. Farinaro, andM.Trevisan, āUricacid metabolism and tubular sodium handling: results froma population-based study,ā Journal of the American MedicalAssociation, vol. 270, no. 3, pp. 354ā359, 1993.
[22] M. Modan, H. Halkin, A. Karasik, and A. Lusky, āElevatedserum uric acidāa facet of hyperinsulinaemia,ā Diabetologia,vol. 30, no. 9, pp. 713ā718, 1987.
[23] T. Nakagawa, H. Hu, S. Zharikov et al., āA causal role foruric acid in fructose-induced metabolic syndrome,ā AmericanJournal of Physiology. Renal Physiology, vol. 290, no. 3, pp. F625āF631, 2006.
[24] S. Baldus, R. Koster, P. Chumley et al., āOxypurinol improvescoronary and peripheral endothelial function in patients withcoronary artery disease,ā Free Radical Biology andMedicine, vol.39, no. 9, pp. 1184ā1190, 2005.
[25] L. Jia, J. Xing, Y. Ding et al., āHyperuricemia causes pancreaticš½-cell death and dysfunction through NF-š B signaling path-way,ā PLoS ONE, vol. 8, no. 10, Article ID e78284, 2013.
[26] D. G. Cook, A. G. Shaper, D. S. Thelle, and T. P. Whitehead,āSerum uric acid, serum glucose and diabetes: relationships ina population study,ā Postgraduate Medical Journal, vol. 62, no.733, pp. 1001ā1006, 1986.
[27] J. Vasquez-Vivar, A. M. Santos, V. B. Junqueira, and O. Augusto,āPeroxynitrite-mediated formation of free radicals in human
plasma: EPR detection of ascorbyl, albumin-thiyl and uric acid-derived free radicals,āBiochemical Journal, vol. 314, pp. 869ā876,1996.
[28] A. P. Robertson, āChronic oxidative stress as a central mech-anism for glucose toxicity in pancreatic islet beta cells indiabetes,ā Journal of Biological Chemistry, vol. 279, no. 41, pp.42351ā42354, 2004.
[29] Y. Zhang, T. Yamamoto, I. Hisatome et al., āUric acid inducesoxidative stress and growth inhibition by activating adeno-sine monophosphate-activated protein kinase and extracellularsignal-regulated kinase signal pathways in pancreatic š½ cells,āMolecular and Cellular Endocrinology, vol. 375, no. 1-2, pp. 89ā96, 2013.
[30] F. J. Nieto, C. Iribarren, M. D. Gross, G. W. Comstock, and R.G. Cutler, āUric acid and serum antioxidant capacity: a reactionto atherosclerosis?ā Atherosclerosis, vol. 148, no. 1, pp. 131ā139,2000.
[31] K. Nyyssonen, E. Porkkala-Sarataho, and J. Kaikkonen, āAscor-bate and urate are the strongest determinants of plasma antiox-idative capacity and serum lipid resistance to oxidation inFinnishmen,āAtherosclerosis, vol. 130, no. 1-2, pp. 223ā233, 1997.