Features of the metabolic syndrome predict higher risk of diabetes and impaired glucose tolerance: a prospective study in Mauritius

Please download to get full document.

View again

of 8
19 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
Features of the metabolic syndrome predict higher risk of diabetes and impaired glucose tolerance: a prospective study in Mauritius
Document Share
Document Transcript
  See discussions, stats, and author profiles for this publication at: https://www.researchgate.net/publication/12345264 Features of the Metabolic Syndrome predicthigher risk of diabetes and impaired glucosetolerance  Article   in  Diabetes Care · October 2000 DOI: 10.2337/diacare.23.9.1242 · Source: PubMed CITATIONS 92 READS 26 6 authors , including: Some of the authors of this publication are also working on these related projects: Screening for type 2 diabetes – validation of a diabetes risk score in Saudi Arabia   View projectDeterminant Factors of Glycemic Control Among type 2 Diabetes Saudi Patients   View projectMaximilian Pangratius de CourtenVictoria University Melbourne 142   PUBLICATIONS   6,444   CITATIONS   SEE PROFILE Jaakko TuomilehtoDanube University Krems 1,525   PUBLICATIONS   122,733   CITATIONS   SEE PROFILE George AlbertiKing's College Hospital NHS Foundation Trust 1,120   PUBLICATIONS   71,807   CITATIONS   SEE PROFILE All content following this page was uploaded by Maximilian Pangratius de Courten on 04 December 2016. The user has requested enhancement of the downloaded file. All in-text references underlined in blue are added to the srcinal documentand are linked to publications on ResearchGate, letting you access and read them immediately.  1242 D IABETES C ARE , VOLUME 23, NUMBER 9, S EPTEMBER 2000 T he more frequent occurrence thanexpected by chance of multiple char-acteristics associated with poor cardio-vascular and other health outcomes such ashyperglycemia, hypertension, and lipidabnormalities has given rise to the concept of the “metabolic syndrome” (1). A recentreview of this syndrome identified 6 com-ponents that were included in most of thedefinitions (fasting or stimulated circulatinginsulin and glucose levels, triglyceride levels,HDL cholesterol levels, blood pressure lev-els, and overall or central obesity) (1). Alsoincluded in some definitions were serumuric acid (2–4) and total or LDL cholesterol(5) levels. Recent data from our group (6),the results of a factor analysis by others (7),and the demonstration of associationsbetween leptin and both insulin sensitivity(8) and hyperinsulinemia (9) argue for lep-tin’s involvement in the metabolic syndrome. The significance of this syndromeregarding the development of cardiovasculardisease is fairly well understood becausemany of its components are well-acceptedcardiovascular disease risk factors, but this isless true regarding the prediction of deterio-ration in glucose tolerance. Although manystudies have examined features of the meta-bolic syndrome in relation to risk of diabetesand impaired glucose tolerance (IGT), fewhave examined the effects of all 6 compo-nents. A 3.5-year follow-up of an elderlyFinnish cohort that considered these 6 com-ponents revealed higher diabetes risk inassociation with presence of IGT or hyper-tension; higher fasting and 2-h plasmainsulin levels, total triglyceride level, waist-to-hip ratio (WHR), and BMI; and lowerHDL cholesterol levels (10). The indepen-dent effects of these correlated factors werenot directly assessed in multivariate models. The importance of assessing whether meta-bolic syndrome components have indepen-dent effects on the risk of this outcome wasdemonstrated by an analysis conductedusing San Antonio Heart Study data (11). Amultivariate model developed to predict dia-betes identified fasting and 2-h glucose lev-els, BMI, HDL cholesterol levels, and pulsepressure as having independent effects onthe risk of diabetes during 8 years of follow-up (11). In this model, fasting insulin levels,triglyceride levels, subscapular-to-tricepsskinfold ratio (a measure of central adipos-ity), and systolic and diastolic blood pressuredid not have independent effects on diabetesrisk, even though they were significantlyrelated to higher risk in an age-, sex-, andethnicity-adjusted analysis. Whether theeffects of these independent variables in From the International Diabetes Institute (E.J.B., M.d.C., P.Z.Z.), Melbourne, Victoria, Australia; the Epi-demiologic Research and Information Center (E.J.B.), Veterans Affairs Puget Sound, Seattle, Washington; theMinistry of Health (P.C.), Port Louis, Mauritius; the Department of Epidemiology and Health Promotion (J.T.),National Public Health Institute, Helsinki, Finland; and the Human Diabetes and Metabolism Research Cen-tre (K.G.M.M.A.), University of Newcastle upon Tyne, Newcastle upon Tyne, U.K.Address correspondence and reprint requests to Edward J. Boyko, MD, MPH, Veterans Affairs MedicalCenter (S-111-GIMC), 1660 S. Columbian Way, Seattle, WA 98108. E-mail: eboyko@u.washington.edu.Received for publication 29 June 1999 and accepted in revised form 17 May 2000. Abbreviations: HOMA %S, insulin sensitivity measured with homeostasis model assessment; IGT,impaired glucose tolerance; RR, relative risk; WHR, waist-to-hip ratio.A table elsewhere in this issue shows conventional and Système International (SI) units and conversionfactors for many substances. Features of the Metabolic SyndromePredict HigherRisk of Diabetes andImpaired GlucoseTolerance A prospective study in Mauritius ORIGINAL ARTICLE OBJECTIVE  —To assess the independent and joint effects of the components of the meta-bolic syndrome, including leptin, which is a recently proposed addition to this syndrome, inpredicting the cumulative incidence of impaired glucose tolerance (IGT) and diabetes amongindividuals with normal glucose tolerance. RESEARCH DESIGN AND METHODS  —This prospective study involved 2,605 res-idents of Mauritius with normal glucose tolerance who were followed for 5 years for IGT ordiabetes onset in relation to total and regional adiposity (BMI, waist-to-hip ratio [WHR]), fast-ing and 2-h 75-g oral glucose load glucose and insulin, total and HDL cholesterol, blood pres-sure, serum uric acid, triglyceride, and leptin levels. RESULTS  —A multivariate logistic regression model adjusted for age, sex, ethnicity, and dia-betes family history showed a significantly higher linear increase in risk of IGT and diabetesin association with the following variables only: fasting glucose (odds ratio 1.89 [95% CI1.51–2.34]), 2-h glucose (1.68 [1.50–1.88]), WHR (1.30 [1.10–1.52]), BMI (1.04 [1.00–1.08]),and serum uric acid (1.37 [1.20–1.57]). However, a nonlinear increase was seen with serumtriglyceride and plasma leptin concentrations. No risk factors resulted in joint effects that weregreater than expected from combining individual effects. CONCLUSIONS  —Metabolic syndrome features independently predict a higher risk of dia-betes or IGT in normoglycemic subjects but in combination confer no higher-than-expectedrisk of these outcomes. At higher concentrations of triglycerides and leptin, risk plateaus andeven declines slightly. Diabetes Care  23:1242–1248, 2000 E DWARD  J. B OYKO , MD , MPH M AXIMILIANDE C OURTEN , MD P AUL Z. Z IMMET , MD P IERROT C HITSON , MBBS  J AAKKO T UOMILEHTO , PHD K. G EORGE M.M. A LBERTI , DPHIL Ep i d em i o l o g y / H ea l th Ser v i ces/ Psy ch o so ci a l Resea r ch  D IABETES C ARE , VOLUME 23, NUMBER 9, S EPTEMBER 2000 1243 Boyko and Associates  combination had effects on diabetes riskthat were greater than expected from com-bining individual effects was not assessed.Few studies have examined in a com-prehensive fashion the effect of most meta-bolic syndrome components and leptinconcentration on the risk of diabetes andIGT in a manner that permitted the assess-ment of the independent effects and testingfor interaction between individual compo-nents. We therefore investigated this issueamong residents of Mauritius with normalglucose tolerance at baseline who were fol-lowed for 5 years for the development of IGT or diabetes. RESEARCH DESIGN AND METHODS Background and subjects Mauritius is a subtropical island located inthe southwestern Indian Ocean about 800km east of Madagascar. The multiethnicpopulation consists of 70% individuals of Indian srcin (54% Hindu and 16% Mus-lim), 2% individuals of Chinese origin, andthe remaining 28% are Creoles predomi-nantly of African and Malagasy ancestrywith some European admixture.All adults 25–74 years of age livingwithin 11 geographically defined areas wereeligible for the 1987 baseline survey as pre-viously described (12,13). Overall response( n  =5,083) consisted of 83% of eligible menand 89% of eligible women. All subjectsalive and willing to participate were eligiblefor the 1992 follow-up survey. Longitudinal1987–1992 data were available for 3,793subjects (74.6%). This analysis focuses onthe development of diabetes and IGTamong individuals who were initially nor-moglycemic (as defined below) in 1987 ( n  =3,467), of whom 2,605 (75.1%) had lon-gitudinal information available from bothsurveys. The survey protocol was reviewedand approved by the Alfred HealthcareGroup Ethics Committee (Melbourne, Aus-tralia). Survey procedures  The following procedures were carried outon all subjects in 1987 as previouslydescribed (12–15). Height was measuredwithout shoes to the nearest centimeter.Weight was measured to the nearest 0.1 kgin light clothes and without shoes. BMI wascalculated as weight in kilograms by heightin meters squared. Waist and hip circum-ferences were measured in duplicate to thenearest 0.5 cm with a measuring tape whilethe subject was standing relaxed and wear-ing 1 layer of light clothing. A third mea-surement was taken if the first 2 were notwithin 2 cm of each another. The mean of the closest 2 measurements was used tocalculate the WHR. The waist measurementwas taken at the midpoint between the iliaccrest and the lower rib margin, and the hipcircumference was taken around the maxi-mum circumference of the buttocks poste-riorly and the symphysis pubis anteriorly.Participants were seated and rested for atleast 5 min before systolic and fifth-phasediastolic blood pressure measurements weretaken in duplicate using a standard mercurysphygmomanometer. Ethnicity was deter-mined by self-report.All subjects not taking diabetes medica-tion had a 2-h 75-g (glucose monohydrate)oral glucose tolerance test after an overnightfast in 1987 and 1992. Diabetes was diag-nosed if subjects reported a history of dia-betes and were taking oral hypoglycemicmedication or insulin or if the fasting plasmaglucose level was  7.0 mmol/l or the 2-hvalue was  11.1 mmol/l (16). Subjects witha fasting plasma glucose level  7.0 mmolbut a 2-h value from  7.7 to  11.1 weredefined as having IGT. Subjects who did notmeet the criteria for either IGT or diabeteswere deemed normoglycemic (16).Fasting and 2-h venous blood sampleswere centrifuged and separated immediately,and plasma glucose was measured on-siteusing Yellow Springs Instruments glucoseanalyzers (Yellow Springs, OH) within 3 h of collection. Serum samples were stored at  20°C and were transported on dry icefrom Mauritius to Newcastle upon Tyne,U.K., in 1987 and later to Melbourne in1996. Serum fasting and 2-h insulin mea-surements were analyzed in Newcastle upon Tyne using a modified method of Soeldnerand Slone (17). The interassay and intra-assay coefficients of variation were 6 and 4%,respectively. Serum uric acid, fasting serumtriglyceride, and HDL cholesterol levels inplasma were measured using manual enzy-matic methods at the central laboratory inMauritius. For quality assurance, every 10thsample was also analyzed in Newcastleupon Tyne. Triglyceride and total choles-terol values were consistently overestimatedacross the distribution and were adjusteddownward using a calculated regressionequation. Leptin was measured in 1996 byradioimmunoassay (Linco, St. Charles, MO). The limit of detection for the assay is 0.5ng/ml in human serum or plasma withacceptable interassay (8.2%) and intra-assay(4.1%) coefficients of variation. A measureof insulin sensitivity was estimated usinghomeostasis model assessment (HOMA%S) (18). Statistical analysis All statistical analyses were performedusing Stata Version 6.0 (College Station, TX). Logistic regression analysis was usedto estimate odds ratios (95% CIs) and P  val-ues for the independent associationsbetween metabolic syndrome componentsand the odds of developing diabetes or IGTcoded as a dichotomous variable (0 =nor-moglycemia, 1 =IGT or diabetes) (19). Abackwards selection algorithm was used toarrive at the best final model. To assesswhether nonlinear effects were presentbetween diabetes or IGT occurrence andmetabolic syndrome components, the riskof the dependent variable was plottedagainst the independent variable dividedinto 20 quantiles, and visual inspectionwas used to assess departures from linear-ity, which were modelled with appropriatetransformations. To assess whether effectmodification existed when multiple com-ponents of the metabolic syndrome werepresent simultaneously, first-order interac-tion terms were tested for significance inlogistic regression models with the thresh-old for rejection of the null hypothesis of no interaction set at P   0.05. This analy-sis permitted formal assessment of whether joint effects of metabolic syndrome com-ponents exceeded the expected combina-tion of individual effects on the incidence of the outcomes of interest. Relative risk (RR)of diabetes or IGT was calculated in crudeand stratified analyses using standardmethods (19). RESULTS  — A total of 2,605 subjectswith normal glucose tolerance in 1987 andfollow-up data in 1992 were available foranalysis. Mean age for study subjects was40.4 years and ranged from a low of 25years to a high of 74 years. Subjectsincluded 48.9% men and 51.1% women.Among these individuals, 356 (13.7%)developed IGT and 159 (6.1%) developeddiabetes during the 5-year follow-up,which gave annual incidence rates of 2.9and 1.3%, respectively.Characteristics of study subjects bywhether diabetes or IGT developed duringfollow-up in univariate analyses are shownin Table 1. Sex, ethnicity, and diabetes fam-ily history were not significantly related todiabetes or IGT odds. Regarding features of   1244 D IABETES C ARE , VOLUME 23, NUMBER 9, S EPTEMBER 2000 Diabetes and IGT risk in the metabolic syndrome  the metabolic syndrome, all measures weresignificantly related to diabetes or IGT oddsin the expected directions. Higher plasmaleptin was significantly related to higherodds of diabetes or IGT. By visual examina-tion, plots of the risk of diabetes or IGT inrelation to triglyceride quantiles appearednonlinear with a plateau followed by a slightdecline at higher levels (Fig. 1 A ). A similarassociation was seen for plasma leptin (Fig.1 B  ). Because leptin levels differ markedly bysex in this and other populations, this rela-tionship was reexamined by sex, but similarassociations were seen (Fig. 1 C  and D  ). Therefore, multivariate models were fit withlog e transformations of these variables inaddition to the linear term (Table 1).Because both the linear and log e transfor-mations of these variables were significantin these models at P   0.05, this findingrejects the hypothesis that a linear associa-tion exists between diabetes or IGT oddsand leptin or triglyceride levels. The relativeodds of diabetes or IGT in relation to quin-tiles of leptin and triglycerides inserted intoeach model shown in Table 1 as a set of dummy variables demonstrated the levelingoff of risk with increasing level, more so forleptin than for triglycerides.Analysis of leptin effects on diabetes orIGT risk was further examined by estimatingthe RR of diabetes or IGT in crude and strat-ified analyses. In crude analysis, the RR(95% CIs) of diabetes or IGT for leptin con-centration above the median value was 1.7(1.4–2.0). A higher RR was seen in men (2.0 Table 1— Baseline characteristics of study subjects in relation to the occurrence of IGT and diabetes at the end of the follow-up period  Normal glucoseDiabetes or IGT vs. normalCharacteristictoleranceIGTDiabetesglucose tolerance n  2,090356159Age (years)39.5 ±11.843.4 ±12.145.3 ±12.33.3 (2.4–4.5)SexMen48.740.770.41.0 (referent)Women51.359.329.61.0 (0.8–1.2)EthnicityIndian (Hindu)54.855.154.71.0 (referent)Indian (Muslim)14.917.411.31.0 (0.8–1.4)African Creole24.921.126.40.9 (0.7–1.2)Chinese5.46.57.61.3 (0.8–1.9)Family history of diabetes20.122.821.41.1 (0.9–1.4)BMI (kg/m 2 )22.8 ±3.824.6 ±4.524.2 ±4.31.5 (1.4–1.7)WHR0.83 ±0.070.85 ±0.070.88 ±0.081.7 (1.5–1.9)Systolic blood pressure (mmHg)122.0 ±16.8127.4 ±19.3131.2 ±19.11.3 (1.2–1.5)Diastolic blood pressure (mmHg)75.3 ±11.278.5 ±11.281.5 ±12.71.4 (1.3–1.5)Serum triglycerides (mmol/l)1.35 ±1.041.50 ±0.931.83 ±1.31See triglycerides model belowPlasma HDL cholesterol (mmol/l)1.30 ±0.341.27 ±0.301.22 ±0.320.8 (0.8–0.9)Serum uric acid (mmol/l)0.34 ±0.080.36 ±0.090.40 ±0.091.5 (1.4–1.7)Fasting plasma insulin (µU/ml)6.48 ±5.538.60 ±6.508.38 ±7.881.4 (1.3–1.5)2-h plasma insulin (µU/ml)37.67 ±34.6254.05 ±49.8148.85 ±50.231.4 (1.3–1.6)HOMA %S114.6 ±108.885.6 ±92.198.9 ±101.30.7 (0.6–0.8)Fasting plasma glucose (mmol/l)5.10 ±0.495.25 ±0.545.51 ±0.571.5 (1.4–1.7)2-h plasma glucose (mmol/l)5.61 ±1.146.39 ±0.986.20 ±1.072.1 (1.9–2.4)Fasting plasma leptin (ng/ml)8.52 ±9.6910.56 ±8.807.52 ±6.86See leptin model belowLeptin model (leptin  log e plasma leptin, 1.7 (1.4–2.1)25th vs. 1st percentile)Leptin quintiles1 (lowest)1.0 (referent)21.0 (0.7–1.4)31.9 (1.3–2.7)42.2 (1.6–3.1)52.6 (1.8–3.6) Triglycerides model (triglycerides  log e serum 2.1 (1.6–2.8)triglycerides, 25th vs. 1st percentile) Triglyceride quintiles1 (lowest)1.0 (referent)21.3 (0.9–1.9)31.8 (1.3–2.5)41.7 (1.2–2.4)52.8 (2.0–3.9) Data are n  , means ±SD, %, or odds ratios (95% CIs). Odds ratios for continuous variables reflect a 1-SD magnitude increase.  D IABETES C ARE , VOLUME 23, NUMBER 9, S EPTEMBER 2000 1245 Boyko and Associates  [1.6–2.5]) than in women (1.5 [1.2–1.8]),but this difference was not significant by theheterogeneity  2 test ( P  =0.07). In analysesstratified by BMI and WHR dichotomized atthe median value (BMI: men 22.5 kg/m 2 ,women 23.0 kg/m 2 ; WHR: men 0.88,women 0.79), RRs associated with leptinconcentration above the median value didnot differ at higher versus lower levels of overall adiposity (BMI less than or equal tothe median 1.3 [0.9–1.7]; BMI greater thanthe median 1.3 [1.0–1.7]) or at higher ver-sus lower levels of WHR (WHR less than orequal to the median 1.5 [1.1–2.0]; WHRgreater than the median 1.4 [1.1–1.7]).All metabolic syndrome features wereinserted into 2 logistic regression models (1containing HOMA %S and the other con-taining insulin levels to avoid collinearity)for the prediction of diabetes or IGT thatalso contained age, sex, ethnicity, and dia-betes family history (Table 2, full model). Inthese full models, the following variableswere no longer significantly related to theoutcome: BMI, HDL cholesterol level,HOMA %S, systolic and diastolic bloodpressure, and fasting and 2-h insulin levels.Using backward stepwise logistical regres-sion on these full models yielded 1 reducedmodel (Table 2) that contained variablesthat independently predict diabetes or IGToccurrence. Nonlinear associations betweenthese continuous variables (except for leptinand triglycerides) and diabetes incidencewere tested using exponential and logarith-mic transformations, but none representedan improvement in fit compared with themodel shown that used simple linear terms. This analysis showed that WHR, fastingand 2-h plasma glucose levels, serum uricacid levels, leptin concentrations, andtriglyceride levels significantly predicted theoccurrence of diabetes or IGT, whereas BMIwas nearly significantly related to this out-come ( P  =0.053). The reduced model was repeated withdiabetes only as the outcome (Table 2).Odds ratios of a similar magnitude wereseen when compared with the diabetes orIGT reduced model except in severalinstances. The odds ratios for serumtriglycerides and the log of this value werecloser to 1, whereas the fasting glucoseodds ratio was greater. Statistical signifi-cance of these associations differed aswould be expected because of the smallernumber of diabetes ( n  =159) versus dia-betes or IGT ( n  =515) outcomes.Interaction was tested by the insertion of first-order interaction terms between all pos-sible pairs ( n  =35) of independent variablesin the reduced model predicting diabetes orIGT. None of these interactions was signifi-cant. The interaction between BMI and lep-tin came closest with a P value of 0.107, but P values for the remaining terms were gen-erally much larger. Also, no significant dif-ference in the effects of the variables in thereduced model predicting diabetes or IGTwere seen according to age, sex, ethnicity, ordiabetes family history. Relative odds werecalculated using the reduced logistic modelcoefficients for different comparisons of lep-tin and triglyceride levels adjusted for allcovariates (Table 2). These comparisonsdemonstrate an increase in the odds of dia-betes or IGT when these levels increase fromnear the minimum value, with leveling off and eventually a small decline in the odds athigher values (Table 2). CONCLUSIONS  —  This analysisdemonstrates that all features of the meta-bolic syndrome that we considered wererelated in the univariate analysis to a higherrisk of worsening of glucose toleranceamong members of this multiethnic study Figure 1— Risk of the combined outcome in relation to quantiles (  n= 20) of serum triglycerides (  A  ) and plasma leptin levels (  B–D  ). Mean triglyceride and leptin levels for each quantile are shown on the x -axis, and risk of diabetes or IGT (DM/IGT) is shown on the y -axis. ADCB All subjectsMen OnlyWomen OnlyAll subjects
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