A diabetes risk score helps identify metabolic syndrome and cardiovascular risk in Indians ? the Chennai Urban Rural Epidemiology Study (CURES-38

Please download to get full document.

View again

of 7
6 views
PDF
All materials on our website are shared by users. If you have any questions about copyright issues, please report us to resolve them. We are always happy to assist you.
Document Description
Aims:  The aims of the study were to compare the recently evolved Indian Diabetes Risk Score (IDRS), in subjects with different grades of glucose intolerance and to evaluate its usefulness as an indicator of cardiovascular risk in Asian Indians, a
Document Share
Document Tags
Document Transcript
  A diabetes risk score helps identify metabolic syndrome andcardiovascular risk in Indians – the Chennai Urban RuralEpidemiology Study (CURES-38) V. Mohan, 1 S. Sandeep, 1 M. Deepa, 1 K. Gokulakrishnan, 1 M. Datta 2 and R. Deepa 1 1 Madras Diabetes Research Foundation & Dr Mohan’s Diabetes Specialities Centre, Chennai, Tamil Nadu, India 2 Department of Epidemiology, The Tamil Nadu Dr M.G.R. Medical University, Chennai, Tamil Nadu, India Aims:  The aims of the study were to compare the recently evolved Indian Diabetes Risk Score (IDRS), in subjectswith different grades of glucose intolerance and to evaluate its usefulness as an indicator of cardiovascular risk inAsian Indians, a high risk group for diabetes and coronary artery disease (CAD). Methods:  ThedataforthepresentstudywereobtainedfromthePhase3(n  ¼  2350,responserate:90.4%)oftheChennaiUrbanRuralEpidemiologyStudy,apopulation-basedstudydoneinChennai,thelargestcityinsouthernIndia.IDRSwasdeveloped based on multiple logistic regression analysis using four simple parameters namely age, abdominal obesity,familyhistoryoftype2diabetesandphysicalactivity.Inallsubjects,familyhistoryofdiabeteswasobtained,anddetailsonphysicalactivitywereassessedusing a validated questionnaire. Subjects with anIDRS of   < 30 was categorized aslowrisk, 30–50 asmediumriskand thosewith  60ashighriskfor diabetes. Biochemical and anthropometricmeasurementswere done using standardized procedures. Minnesota coding was used to grade 12-lead electrocardiogram. Results:  The mean IDRS increased significantly with worsening glucose intolerance [normal glucose tolerance (NGT)subjects: 48    17, impaired glucose tolerance (IGT): 57    16, newly diagnosed diabetics (NDD): 61    15 and knowndiabetics (KD): 68    12; p for trend  < 0.001]. Among NGT group, the prevalence of cardiovascular risk factorsincreased progressively in low-, medium- to high-risk score groups; hypertension: 9.4, 22.1 and 38.2% (p for trend: < 0.001), hypertriglyceridemia: 8.8, 19.9 and 25.3% (p for trend:  <  0.001), hypercholesterolemia: 7.2, 20.3 and 34.9%(p for trend:  <  0.001) and metabolic syndrome: 1.8, 14.6 and 30.3% (p for trend:  <  0.001), respectively. Theprevalence of CAD was also significantly higher in individuals with high risk compared with those with low risk(p  ¼  0.030) and the medium risk (p  ¼  0.050) in the NGT group. Conclusions:  The results suggest that in Asian Indians, (i) the diabetes risk score increases with increasing glucoseintolerance, and (ii) it can serve as an effective indicator of metabolic syndrome and cardiovascular risk even amongsubjects with NGT. Keywords: Asian Indians, cardiovascular risk factors, coronary artery disease, diabetes, Indian diabetes risk score, metabolic syndrome Received 24 February 2006; revised version accepted 19 March 2006 Introduction Cardiovascular mortality with particular reference tocoronary artery disease (CAD) has increased worldwide[1]. It is predicted that India would lead the world withthe highest rates of cardiovascular deaths in the forth-coming years [2]. In addition, India is already the largest doi: 10.1111/j.1463-1326.2006.00612.x Correspondence: Dr Viswanathan Mohan, Chairman and Chief of Diabetology, Madras Diabetes Research Foundation & Dr Mohan’s DiabetesSpecialities Centre, 4, Conran Smith Road, Gopalapuram, Chennai 600086, Tamil Nadu, India. E-mail: vhans@vsnl.net ORIG INAL ARTICLE #  2006 The AuthorsJournal Compilation #  2006 Blackwell Publishing Ltd  Diabetes, Obesity and Metabolism,  9,  2007, 337–343  j  337  contributor of diabetes, with over 32 million diabeticindividuals in 2000 [3] of whom over half remainundiagnosed [4]. The prevalence of impaired glucosetolerance (IGT), a known risk factor for cardiovascularmortality [5], is also high in India [6,7]. With the pre-valence of diabetes increasing rapidly even in ruralIndia [8], early detection of subjects with high risk fordiabetes and cardiovascular disease using a simple butrobust tool would be of great clinical significance andhelp make screening programs more cost effective.In this context, we recently evolved the IndianDiabetes Risk Score (IDRS), a simplified risk score foridentifying undiagnosed diabetic subjects using foursimple parameters – age, waist circumference, familyhistory of diabetes and physical activity [9]. IDRS is aninexpensive tool for screening undiagnosed diabetes,and it requires minimum time and effort. A score   60was found to have optimum sensitivity and specificityfor detecting undiagnosed diabetes [9].The present study was undertaken with the followingaims: first, to see how IDRS performs in individualswith different grades of glucose intolerance, namelynormal glucose tolerance (NGT), IGT, newly diagnoseddiabetes (NDD) and known diabetes (KD); second, tolook at the prevalence of cardiovascular risk factorsamong subjects with NGT with different IDRS. Methods The data for the present study were obtained from theChennai Urban Rural Epidemiology Study (CURES), themethodology of which is detailed elsewhere [10].Chennai (formerly Madras), the largest city in SouthernIndia and the fourth largest in India, has a population of 4.2 million. The Chennai corporation has dividedChennai into 10 zones and 155 wards. The samplingfor CURES was based on the model of systematic ran-dom sampling, wherein the 155 wards, 46 were selectedto represent all the 10 zones. Twenty-six thousand andone individuals were selected from these 46 wards forthe Phase 1 of CURES. Phase 2 focused on the self-reported diabetic subjects identified in Phase 1, whilePhase 3 recruited every 10th subject (n  ¼  2600)screened in Phase 1. Phase 3 had a response rate of 90.4% (i.e. 2350/2600 subjects participated). These indi-viduals were invited to our centre for an oral glucosetolerance test (OGTT) (KD subjects underwent fastingand postprandial glucose tests). In all subjects, familyhistory of diabetes was obtained, and details on physicalactivity were assessed using a validated questionnaire[10]. Waist measurements in centimetres were obtainedusing standardized techniques [10]. Plasma glucose (PG)levels (glucose oxidase-peroxidase) were measured onHitachi-912 Autoanalyser (Hitachi, Mannheim,Germany) using kits supplied by Roche Diagnostics(Mannheim, Germany). Diagnosis of diabetes was basedon WHO Consulting Group criteria, that is 2-h venousplasma glucose  11.1 mmol/l [11]. Those who were con-firmed by OGTT to have 2-h PG value  11.1 mmol/l, butdid not have any prior history of diabetes, were labelledas ‘newly detected diabetic subjects’ (NDD). IGT wasdefined as 2-h PG ranging from 7.8 to 11.0 mmol/l.The IDRS was developed based on multiple logisticregression model using four simple parameters namelyage, abdominal obesity, family history of diabetes andphysical activity as described elsewhere [9]. The infor-mation for these risk factors was obtained based on foursimple questions and one anthropometric measurementnamely waist circumference. Subjects with an IDRSvalue of   < 30 was categorized as low risk, those between30 and 50 as medium risk and those with   60 as highrisk for diabetes.CAD was diagnosed based on a history of documentedmyocardial infarction and/or medical therapy (aspirinor nitrates), revascularization for CAD and/or electrocar-diographic changes suggestive of ST-segment depres-sion and/or Q-wave changes.Serum cholesterol (cholesterol oxidase-peroxidase-amidopyrine method) serum triglycerides (glycerolphosphate oxidase-peroxidase-amidopyrine method)and high-density lipoprotein (HDL) cholesterol (directmethod-polyethylene glycol-pretreated enzymes) weremeasured using Hitachi-912 Autoanalyser (Hitachi).The intra- and interassay coefficient of variation for the biochemical assays ranged from 3.1 to 7.6%. Low-density lipoprotein (LDL) cholesterol was calculatedusing the Friedewald formula [12]. Glycated haemoglo- bin (HbA1C) was estimated by high-pressure liquidchromatography using the Variant machine (Bio-Rad,Hercules, CA, USA). The intra- and interassay coeffi-cient of variation of HbA1C was  < 10%.Serum insulin concentration was estimated usingenzyme-linked immunosorbent assay (Dako, Glostrup,Denmark). The intra-assay and the interassay coefficientof variation for insulin assay were 5.7 and 8.9%, respec-tively, and the lower detection limit was 0.5  m IU/ml.Hypercholesterolemia (serum cholesterol  5.2 mmol/l),hypertriglyceridemia (serum triglycerides   1.7 mmol/l)and low HDL levels (males: HDL cholesterol < 1.0 mmol/l, females: HDL cholesterol  < 1.3 mmol/l)were diagnosed based on ATPIII guidelines [10].Insulin resistance was calculated using the homeostasismodel assessment (HOMA-IR) using the formula, fastinginsulin ( m IU/ml)    fasting PG (mmol/l)/22.5 [13]. OA  Diabetes risk score and cardiovascular risk  V. Mohan  et al  . 338  j  Diabetes, Obesity and Metabolism,  9,  2007, 337–343 #  2006 The AuthorsJournal Compilation # 2006 Blackwell Publishing Ltd  Metabolic syndrome was diagnosed based on modi-fied ATPIII guidelines [14], if three or more of the fol-lowing were present: abdominal obesity (definition of abdominal obesity was modified using Asia-PacificWHO guidelines as waist circumference   90 cm formales and   80 cm for females [15], hypertension (sub-jects who were on antihypertensive medication and/orhad systolic blood pressure   130 mmHg and/or diasto-lic blood pressure   85 mmHg), glucose intolerance(fasting PG   6.1 mmol/l), hypertriglyceridemia (fastingtriglycerides   1.7 mmol/l) or low HDL levels (males:HDL cholesterol  < 1.0 mmol/l, females: HDL cholesterol < 1.3 mmol/l). Statistical Analysis Student’s t  -testorone-way ANOVA (withTukey’sHSD)wasused to compare groups for continuous variables, and  w 2 test or Fisher’s Exact test as appropriate was used tocompare proportions. Because diabetes is a cardiovascu-lar risk equivalent, we restricted the analysis on associa-tion of IDRS with cardiovascular risk factors to subjectswith NGT. Spearman correlation analysis was used todetermine the relationship between IDRS and cardiovas-cularriskfactors.AllanalysesweredoneusingWindows- based  SPSS  statistical package (Version 10.0, Chicago, IL,USA), and p-values  < 0.05 were taken as significant. Results Table 1 summarizes the general characteristics of thestudy population. Subjects with IGT, NDD and KDwere older than those with NGT. Body mass index(BMI) (p for trend  < 0.001), waist circumference (p fortrend  < 0.001), systolic (p for trend  < 0.001) and diastolic blood pressure (p for trend  < 0.001), serum cholesterol(p for trend  < 0.001), triglycerides (p for trend  < 0.001)and LDL cholesterol (p for trend  < 0.001) increased withthe increase in severity of glucose intolerance.The mean IDRS increased significantly with worsen-ing glucose intolerance (NGT, 48    17; IGT, 57    16;NDD, 61    15; KD, 68    12; and p for trend  < 0.001).The percentage of individuals with values   60 (thehigh-risk group) increased significantly with theincreasing degrees of glucose intolerance, that is 37%of the NGT, 57% of IGT, 73% of NDD and 88% of KDindividuals, respectively.Table 2 summarizes the clinical and biochemical char-acteristics in the NGT group classified based on the IDRSscores. Overall, among subjects with NGT, 28.2% hadIDRS  < 30, 34.9% had IDRS between 30 and 50 and37% had IDRS   60. With increasing risk scores, therewas increase in BMI (p-value for trend  < 0.001), systolic blood pressure (p-value for trend: 0.007), diastolic bloodpressure (p-value for trend:  <  0.001), fasting PG (p-valuefor trend: 0.019), 2-h PG (p-value for trend: 0.004), total Table 1  General characteristics of the different study groups ParametersNormal glucose tolerance(n = 1736)Impaired glucose tolerance(n = 249)Newly diagnosed diabetes(n = 222)Known diabetes(n = 143) Age (year) 37    12 43    13* 45    12* 52    11* y ‡ BMI (kg/m 2 ) 22.6    4.1 24.3    3.6* 24.3    3.1* 24.8    3.6*Systolic BP (mmHg) 116    16 126    20* 128    22* 128    18*Diastolic BP (mmHg) 74    10 78    12* 78    12* 78    10*Fasting plasma glucose(mmol/l)4.7    0.4 5.2    0.8* 8.6    3.3* y  8.6    3.8* y 2-h plasma glucose (mmol/l) 5.6    1.0 8.9    1.0* 15.8    4.0* y  –Serum cholesterol (mmol/l) 4.5    0.9 4.94    1.1* 5.1    0.9* 5.1    0.9*Triglycerides (mmol/l) 1.3    0.8 1.6    1.1* 2.1    1.4* y  1.9    1.1* y HDL cholesterol (mmol/l) 1.1    0.3 1.1    0.3* 1.0    0.2* 1.1    0.3*LDL cholesterol (mmol/l) 2.8    0.8 3.1    0.9* 3.1    0.8* 3.1    0.9*Glycated haemoglobin (%) 5.6    0.55 6.0    0.72* 8.4    2.1* y  8.7    2.3* y Fasting insulin (pmol/l) 60.0    39.3 92.8    45.2* 78.8    53.7* 76.5    61.8*2-h insulin (pmol/l) 377.5    310.9 677.8    395.8* 493.9    381.0* 332.4    309.5 y IDRSMean    SD 47.8    16.9 56.6    16.3 60.7    15.1 67.3    12.2Score    60 37.0% 57.4% 72.5% 87.4% BMI, body mass index; HDL, high-density lipoprotein; IDRS, Indian diabetes risk score; LDL, low-density lipoprotein.*p-value  <  0.05 compared with normal glucose tolerance. y p-value  <  0.05 compared with impaired glucose tolerance.‡p-value  <  0.05 compared with newly diagnosed diabetics. V. Mohan  et al  .  Diabetes risk score and cardiovascular risk  OA #  2006 The AuthorsJournal Compilation #  2006 Blackwell Publishing Ltd  Diabetes, Obesity and Metabolism,  9,  2007, 337–343  j  339  cholesterol (p-value for trend:  <  0.001), triglycerides(p-value for trend:  <  0.001), LDL cholesterol (p-valuefor trend:  <  0.001), glycated haemoglobin (p-value fortrend:  <  0.001), fasting insulin (p-value fortrend:  <  0.001), 2-h insulin (p-value for trend: 0.004)and HOMA-IR (p-value for trend:  <  0.001).Prevalence of cardiovascular risk factors increasedwith increasing IDRS in the NGT group. In individualswith IDRS scores  < 30, 30–50, and   60, the prevalenceof the following were, hypertension: 9.4, 22.1 and 38.2%(p for trend:  < 0.001), hypertriglyceridemia: 8.8, 19.9and 25.3% (p for trend:  < 0.001) and hypercholesterole-mia: 7.2, 20.3 and 34.9% (p for trend:  < 0.001).Figure 1 shows that the prevalence of metabolic syn-drome increased with IDRS;  < 30: 1.8%, 30–59: 14.6%and   60: 30.3% (p for trend:  < 0.001).The prevalence of CAD was 0.6, 0.8 and 2.2% in thelow-, medium- and high-risk IDRS groups, respectively(figure 2). The prevalence of CAD in the high-risk groupwas significantly higher compared with the low-riskgroup (p  ¼  0.030) and the medium-risk group(p  ¼  0.050). Table 2  Clinical and biochemical characteristics of subjects with normal glucose tolerance with low-, medium- andhigh-IDRS ParameterLow risk (IDRS  < 30)(n = 489)Medium risk (IDRS 30–50)(n = 605)High risk (IDRS  ‡   60)(n = 642) p for trend Age (year) 27    5 35    9 47    12 –BMI (kg/m 2 ) 20.6    3.1 22.9    4.1* 23.9    4.1* y  < 0.001Systolic BP (mmHg) 110    12 114    16* 122    18* 0.007Diastolic BP (mmHg) 68    10 72    10* 76    10*  < 0.001Fasting plasma glucose (mmol/l) 4.6    0.4 4.7    0.4 4.8.    0.5* 0.0192-h plasma glucose (mmol/l) 5.3    1.1 5.6    1.1* 5.7    1.1* 0.004Serum cholesterol (mmol/l) 4.4    0.8 4.6    0.8* 4.6    0.9*  < 0.001Triglycerides (mmol/l) 1.0    0.6 1.4    0.7* 1.4    0.8*  < 0.001HDL cholesterol (mmol/l) 1.1    0.2 1.1    0.3 1.1    0.3 0.305LDL cholesterol (mmol/l) 2.5    0.7 2.6    0.7* 3.1    0.8*  < 0.001Glycated haemoglobin (%) 5.36    0.53 5.57    0.55* 5.75    0.60*  < 0.001Fasting insulin (pmol/l) 52.7    34.1 59.2    44.0 63.7    40.5*  < 0.0012-h insulin (pmol/l) 328.9    278.0 358.9    304.2 411.8    324.3* 0.004HOMA-IR 1.56    1.09 1.77    1.36 1.96    1.28* y  < 0.001 BMI, body mass index; HDL, high-density lipoprotein; IDRS, Indian diabetes risk score; LDL, low-density lipoprotein.*p-value  <  0.05 compared with low risk. y p-value  <  0.05 compared to medium risk group. 1.8%05101520253035Low risk(IDRS < 30)Medium risk(IDRS 30–50)High risk(IDRS ≥  60)30.3%14.6% * † * Fig. 1  Prevalence of metabolic syndrome with increasing Indian Diabetes Risk Score (IDRS) in subjects with normal glucosetolerance. *p-value  < 0.001 compared with low risk.  y p-value  < 0.001 compared with medium risk. OA  Diabetes risk score and cardiovascular risk  V. Mohan  et al  . 340  j  Diabetes, Obesity and Metabolism,  9,  2007, 337–343 #  2006 The AuthorsJournal Compilation # 2006 Blackwell Publishing Ltd  Table 3 summarizes the results of the correlation ana-lysis of IDRS with cardiovascular risk factors in subjectswith NGT. Systolic blood pressure (r  ¼  0.374,p  <  0.001), diastolic blood pressure (r  ¼  0.331,p  <  0.001), total cholesterol (r  ¼  0.325, p  <  0.001),serum triglycerides (r  ¼  311, p  <  0.001), LDL choles-terol (r  ¼  0.260, p  <  0.001) and HOMA-IR (r  ¼  0.252,p  <  0.001) had significant correlation with IDRS. Discussion The rising trend in diabetes and cardiovascular diseasein India necessitates early detection and prevention.Several studies from the west [16,17] and the Indiansubcontinent [18] have formulated diabetes risk scoresfor detecting undiagnosed diabetes. Most of these scoreshave used several questions, inclusive of dietary patternand use of antihypertensive drugs [16,17] while othershad included both waist circumference and BMI in thescore [18].We have earlier reported on the formulation of theIDRS [9]. On the basis of logistic regression modelsused to derive the score, a cut off value of    60 hadoptimum sensitivity (72.5%) and specificity (60.1%)for determining diabetes, while the area under theROC curve was 0.698 [9]. Thus the IDRS was foundto be useful in detecting undiagnosed diabetes.However, its association with CAD, metabolic syn-drome and cardiovascular risk factors has not beenexamined so far.Because it is well known that diabetes is a cardiovas-cular risk equivalent, and subjects with IGT are at highrisk for cardiovascular disease, for examining the asso-ciation of IDRS with cardiovascular risk, we restrictedour analysis to subjects with NGT. The results suggestthat the individuals with ‘medium-’ and ‘high-risk’ hadsignificantly higher prevalence of cardiovascular riskfactors compared with the ‘low-risk’ group. The preva-lence of CAD, hypertension, hypercholesterolemia andhypertriglyceridemia increased significantly withincreasing IDRS among individuals with NGT. Theseresults indicate that subjects with high IDRS values areat risk for cardiovascular disease.Because it is well established that CAD can be pre-vented by interventions like lifestyle modification, useof statins, fibrate or antihypertensive therapy [19–22],it is essential to identify subjects with cardiovascularrisk at the earliest. Studies from the west have shownthat the metabolic syndrome as defined using the NCEP 00.511.522.5Low risk (IDRS   <   30)Medium risk (IDRS 30–50)High risk (IDRS   ≥   60) * †0.6%0.8%2.2% Fig. 2  Prevalence of coronary artery disease with increasing Indian Diabetes Risk Score (IDRS) in subjects with normalglucose tolerance. p-value 0.030 compared with low risk.  y p-value 0.050 compared with medium risk. Table 3  Spearman correlation analysis of IDRS with cardio-vascular risk factors Parameters r p Systolic blood pressure 0.374  < 0.001Diastolic blood pressure 0.331  < 0.001Total cholesterol 0.325  < 0.001Serum triglycerides 0.311  < 0.001HDL cholesterol 0.021 0.319LDL cholesterol 0.260  < 0.001HOMA-IR 0.252  < 0.001 HDL, high-density lipoprotein; IDRS, Indian diabetes risk score;LDL, low-density lipoprotein. V. Mohan  et al  .  Diabetes risk score and cardiovascular risk  OA #  2006 The AuthorsJournal Compilation #  2006 Blackwell Publishing Ltd  Diabetes, Obesity and Metabolism,  9,  2007, 337–343  j  341
Similar documents
View more...
Search Related
We Need Your Support
Thank you for visiting our website and your interest in our free products and services. We are nonprofit website to share and download documents. To the running of this website, we need your help to support us.

Thanks to everyone for your continued support.

No, Thanks