Cardiovascular and Metabolic Risk BRIEF REPORT

Additional Use of for Diagnosis of Type 2 in People Undergoing Coronary Angiography Reveals a Subgroup at Increased Cardiovascular Risk

1,2 5 GUENTHER SILBERNAGEL, MD BERNHARD R. WINKELMANN, MD diagnosis as per the ADA 2009 definition 1,3 6 MARCUS E. KLEBER, PHD BERNHARD O. BOEHM, MD 4 3,4,7 are at increased risk of death from any cause TANJA B. GRAMMER, MD WINFRIED MÄRZ, MD and from cardiovascular diseases (10,11).

RESEARCH DESIGN AND OBJECTIVEdTo study the prognosis of people with newly diagnosed as per d fi METHODS LURIC is a cross-sectional the American Diabetes Association (ADA) 2010 de nition but without diabetes as per the ADA and prospective clinical trial that was de- 2009 definition. signed to investigate cardiovascular risk RESEARCH DESIGN AND METHODSdA total of 2,002 participants of the Ludwigshafen factors. A total of 3,316 white subjects Risk and Cardiovascular Health (LURIC) study without a history of diabetes were studied. were recruited between July 1997 and January 2000 at the Ludwigshafen Heart RESULTSdDuring the follow-up of a mean duration 6 SD of 7.7 6 2.0 years, 346 people died fi n Center in southwestern Germany (10,11). (202 cardiovascular deaths). Subjects with type 2 diabetes as per the ADA 2009 de nition ( = All participants underwent coronary angi- 468) had significantly increased all-cause and cardiovascular mortality compared with people without diabetes as per the ADA 2010 definition (both P # 0.003). Subjects with type 2 diabetes ography. The precise inclusion/exclusion as per the ADA 2010 definition but without diabetes as per the ADA 2009 definition (n =150) criteria have been previously described were at significantly increased risk to die of cardiovascular diseases (P =0.029). (10,11). For the present analyses, subjects with known diabetes or incomplete deter- CONCLUSIONSdUse of the ADA 2010 diabetes definition may be instrumental in improv- mination of the glucometabolic pheno- ing cardiovascular risk stratification in people undergoing coronary angiography. type (missing 75-g oral glucose tolerance test despite fasting glucose ,126 mg/dL) were additionally ruled out. Information on the cause of death was missing for ccording to the 2009 guidelines of Of particular interest, recent data from 11 decedents. These people were ex- Athe American Diabetes Association the Risk in Communities cluded when data on cardiovascular (ADA), subjects with increased fast- (ARIC) study and the Ludwigshafen Risk mortality were analyzed. The study was ing glucose ($126 mg/dL) and/or post- and Cardiovascular Health (LURIC) approved by the ethics committee at the challenge glucose ($200 mg/dL) are study have shown that glycated hemoglo- Ärztekammer Rheinland-Pfalz and was diagnosed with diabetes (1). Using the bin is a better predictor for all-cause and conducted in accordance with the Declara- ADA 2010 criteria, subjects with isolated cardiovascular mortality than fasting glu- tion of Helsinki. Informed written consent elevation of glycated hemoglobin $6.5% cose (9,10). was obtained from all participants (10). Di- (fasting glucose ,126 mg/dL, postchal- The objective of the present work in abetes was diagnosed according to the lenge glucose ,200 mg/dL) are also con- 2,002 LURIC participants was to analyze 2009 and 2010 criteria of the ADA (1,2). sidered diabetic individuals (2). whether subjects with newly diagnosed The follow-up for all-cause and cardiovas- Glycated hemoglobin has been asso- type 2 diabetes as per the ADA 2010 de- cular mortality had a mean duration 6 SD ciated with macrovascular disease (3–8). finition who would not have received the of 7.7 6 2.0 years. ccccccccccccccccccccccccccccccccccccccccccccccccc Laboratory analyses From the 1Ludwigshafen Risk and Cardiovascular Health Study Nonprofit LLC, Freiburg, Germany; the The laboratory methods have been reported 2Division of Endocrinology, Diabetology, Nephrology, Vascular Disease, and Clinical Chemistry, De- previously (10,11). Glucose was measured partment of Internal Medicine, Eberhard-Karls-University Tübingen, Tübingen, Germany; the 3Synlab enzymatically on a Hitachi 717 analyzer Center of Laboratory Diagnostics Heidelberg, Heidelberg, Germany; the 4Mannheim Institute of Public (Roche, Mannheim, Germany). Glycated Health, Social and Preventive Medicine, Mannheim Medical Faculty, University of Heidelberg, Mannheim, hemoglobinwasmeasuredwithanimmu- Germany; the 5Cardiology Group Frankfurt-Sachsenhausen, Frankfurt, Germany; the 6Division of Endo- crinology, Department of Internal Medicine, Ulm University, Ulm, Germany; and the 7Clinical Institute of noassay (hemoglobin A1c UNIMATE 5; Medical and Chemical Laboratory Diagnostics, Medical University of Graz, Graz, Austria. Hoffmann-LaRoche, Grenzach-Whylen, Corresponding author: Winfried März, [email protected]. Germany). Received 2 June 2011 and accepted 15 August 2011. DOI: 10.2337/dc11-1046 Statistical analysis © 2011 by the American Diabetes Association. Readers may use this article as long as the work is properly cited, the use is educational and not for profit, and the work is not altered. See http://creativecommons.org/ The baseline clinical and biochemical licenses/by-nc-nd/3.0/ for details. characteristics are presented for three care.diabetesjournals.org DIABETES CARE 1 Diabetes Care Publish Ahead of Print, published online September 12, 2011 Diabetes categorization and mortality groups: group A, subjects without diabe- Table 1dBaseline characteristics according to diabetes definition tes as per the ADA 2010 definition (with- out diabetes); group B, subjects with type fi No diabetes 2 diabetes as per the ADA 2010 de nition as per ADA 2010 T2DM ADA T2DM ADA but without diabetes as per the ADA 2009 definition 2010 2009 P value* definition (T2DM ADA 2010); and group C, subjects with type 2 diabetes as per the n 1,384 150 468 d ADA 2009 definition (T2DM ADA 2009). Male sex 1,012 (73.1) 113 (75.3) 342 (73.1) 0.839 Categorical data are expressed as numbers Age (years) 60.6 6 10.8 62.7 6 9.2 64.6 6 9.5 ,0.001 and percentages. In the case of continu- Fasting glucose (mg/dL) 98 6 10 103 6 11 125 6 28 ,0.001 ous variables, we report means with SDs Glucose 2 h (mg/dL)† 125 6 33 137 6 30 233 6 65 ,0.001 or medians with interquartile ranges. Fasting insulin (mU/L) 8 (6–12) 9 (6–12) 12 (8–21) ,0.001‡ 6 6 6 , P values for differences in baseline char- Hemoglobin A1c (%) 5.7 0.4 6.8 0.5 6.5 1.0 0.001 acteristics among the three groups were BMI (kg/m2) 27.1 6 3.8 28.0 6 4.3 28.2 6 4.0 ,0.001 calculated with the x2 test for categorical Waist circumference (cm)x 97 (12) 101 (11) 101 (11) ,0.001 data and with ANOVA for continuous Systemic hypertension 920 (66.5) 112 (74.7) 379 (81.0) ,0.001 variables. Triglycerides and insulin were Blood lipid level (mg/dL) transformed logarithmically before being Total cholesterol 196 6 38 189 6 37 193 (40) 0.073 used in parametric statistical procedures. LDL cholesterol 120 6 35 117 6 31 116 6 34 0.097 The Cox proportional hazards model was HDL cholesterol 40 6 11 38 6 10 37 6 10 ,0.001 used to test the relationships of the three Triglycerides 137 (103–190) 140 (108–187) 158 (120–222) ,0.001‡ groups with all-cause and cardiovascular Glomerular filtration rate mortality. Two predefined models of (mL/min/1.73 m2)846 17 80 6 18 81 6 18 ,0.001 adjustment were used (model 1: univari- Smoking 0.074 ate; model 2: adjusted for sex, age, BMI, Never 501 (36.2) 48 (32.0) 153 (32.7) d hypertension, smoking, glomerular fil- Former smoker 597 (43.1) 64 (42.7) 232 (49.6) d tration rate, triglycerides, LDL choles- Current smoker 286 (20.7) 38 (25.3) 83 (17.7) d terol, and HDL cholesterol). The results are presented as hazard ratios with 95% (50% stenosis) 869 (62.8) 102 (68.0) 344 (73.5) ,0.001 CIs. All statistical tests were two-sided Medication use and P , 0.05 was considered significant. b-Blocker 889 (64.2) 88 (58.7) 303 (64.7) 0.369 The SPSS 15.0 statistical package (SPSS ACE inhibitor 667 (48.2) 78 (52.0) 258 (55.1) 0.031 Inc., Chicago, IL) was used. Calcium antagonist 180 (13.0) 19 (12.7) 87 (18.6) 0.010 Diuretic 250 (18.1) 43 (28.7) 166 (35.5) ,0.001 RESULTSdThe clinical and biochemi- Statin 656 (47.4) 75 (50.0) 234 (50.0) 0.561 cal characteristics of the study partici- Acetyl salicylic acid 983 (71.0) 118 (78.7) 322 (68.8) 0.068 pants and data on mortality are shown in Mortality Table 1. A total of 346 (17.3%) deaths All-cause death 194 (14.0) 30 (20.0) 122 (26.3) d occurred during the follow-up. Among Cardiovascular death|| 111 (8.1) 22 (14.8) 69 (14.9) d these, 202 (58.4%) were accounted for Data are n (%) in cases of categorical data and means 6 SDs or medians (25th–75th percentiles) in cases of by cardiovascular diseases. Compared continuous variables. *x2 test and ANOVA for categorical and continuous data, respectively. †n = 1,384/150/ ‡ xn n with subjects without diabetes, people 238. ANOVA of logarithmically transformed values. = 1,363/149/461. || = 1,378/149/464. with T2DM ADA 2009 (hazard ratio 2.02 [95% CI 1.61–2.53]; P , 0.001) and those with T2DM ADA 2010 (1.54 remained significant after multivariate ad- ADA 2009 and those with T2DM ADA [1.05–2.26]; P = 0.028) had increased justment for both subjects with T2DM 2010. In agreement, subjects with T2DM all-cause mortality. After multivariate ad- ADA 2009 (1.62 [1.18–2.21]; P = 0.003) ADA 2009 and people with T2DM ADA justment, this association remained sig- and subjects with T2DM ADA 2010 (1.67 2010 had a similar cardiovascular risk nificant for individuals with T2DM ADA [1.05–2.65]; P = 0.029). There was no factor profile. 2009 (1.62 [1.28–2.04]; P , 0.001) but significant difference in cardiovascular There is broad evidence that type 2 turned insignificant for those with T2DM mortality between subjects with T2DM diabetes will increase the risk of cardio- ADA 2010 (1.34 [0.91–1.97]; P = 0.141). ADA 2009 and subjects with T2DM vascular death (12–14). Thus far, only There was no significant difference in all- ADA 2010 (P = 0.894, model 2). subjects with elevated fasting or postchal- cause mortality between subjects with lenge glucose were diagnosed with diabe- T2DM ADA 2009 and people with CONCLUSIONSdThe current study tes (1). However, fasting and postchallenge T2DM ADA 2010 (P = 0.360, model 2). shows that LURIC participants with glucose have high intraindividual variabil- Compared with subjects without diabe- newly diagnosed T2DM ADA 2010 have ity (15). Hence, the sensitivity of these tests tes, subjects with T2DM ADA 2009 increased cardiovascular mortality com- to select subjects with disturbance of glu- (1.99 [1.48–2.69]; P , 0.001) and those pared with those without diabetes. Of im- cose metabolism is not optimal. with T2DM ADA 2010 (1.98 [1.25–3.13]; portance, there were no differences in Our data support that the use of P = 0.003) more frequently died of car- mortality rates for cardiovascular disease glycated hemoglobin at a cut point of diovascular diseases. These associations between LURIC participants with T2DM $6.5% as an additional criterion for the

2 DIABETES CARE care.diabetesjournals.org Silbernagel and Associates diagnosis of diabetes may improve car- 9. Selvin E, Steffes MW, Zhu H, et al. Gly- diovascular risk stratification in subjects References cated hemoglobin, diabetes, and cardio- 1. American Diabetes Association. Diagnosis referred for coronary angiography. Future fi vascular risk in nondiabetic adults. N Engl and classi cation of diabetes mellitus. Di- J Med 2010;362:800–811 studies should attempt to answer the abetes Care 2009;32(Suppl. 1):S62–S67 question whether measurement of glyca- 10. Silbernagel G, Grammer TB, Winkelmann 2. American Diabetes Association. Diagnosis BR, Boehm BO, März W. Glycated hemo- ted hemoglobin will obviate the need for and classification of diabetes mellitus. Di- globin predicts all-cause, cardiovascular, oral glucose tolerance testing. abetes Care 2010;33(Suppl. 1):S62–S69 and cancer mortality in people without a 3. Khaw KT, Wareham N, Bingham S, Luben R, history of diabetes undergoing coronary Welch A, Day N. Association of hemoglo- angiography. Diabetes Care 2011;34: d bin A1c with cardiovascular disease and Acknowledgments B.O.B. has received grants 1355–1361 mortality in adults: the European pro- from the Deutsche Forschungsgemeinschaft 11. Winkelmann BR, Marz W, Boehm BO, spective investigation into cancer in Nor- (GrK 1041) and the Centre of Excellence et al.; LURIC Study Group (Ludwigshafen “ ” folk. Ann Intern Med 2004;141:413–420 Metabolic Diseases Baden-Wuerttemberg. Risk and Cardiovascular Health). Ratio- The LURIC study has received funding through 4. Gerstein HC, Swedberg K, Carlsson J, d et al.; CHARM Program Investigators. The nale and design of the LURIC study a the 6th Framework Program (integrated project resource for functional genomics, phar- Bloodomics, Grant LSHM-CT-2004-503485) hemoglobin A1c level as a progressive risk macogenomics and long-term prognosis and the 7th Framework Program (integrated factor for cardiovascular death, hospitali- of cardiovascular disease. Pharmacoge- project Atheroremo, Grant 201668) of the zation for heart failure, or death in patients nomics 2001;2(Suppl. 1):S1–S73 European Union. with chronic heart failure: an analysis of 12. Haffner SM,Lehto S,RönnemaaT,Pyörälä K, No potential conflicts of interest relevant to the Candesartan in Heart failure: Assess- Laakso M. Mortality from coronary heart this article were reported. ment of Reduction in Mortality and Mor- G.S. performed the statistical analysis and bidity (CHARM) program. Arch Intern Med disease in subjects with type 2 diabetes and wrote the manuscript. M.E.K. performed the 2008;168:1699–1704 in nondiabetic subjects with and without 5. Brewer N, Wright CS, Travier N, et al. prior myocardial infarction. N Engl J Med statistical analysis. T.B.G. contributed to the – interpretation of results and reviewed the man- A New Zealand linkage study examining 1998;339:229 234 uscript. B.R.W. designed the study. B.O.B. and the associations between A1C concentra- 13. Barr EL, Zimmet PZ, Welborn TA, et al. W.M. designed the study and wrote the manu- tion and mortality. Diabetes Care 2008; Risk of cardiovascular and all-cause mor- script. All authors have read and approved the 31:1144–1149 tality in individuals with diabetes mellitus, manuscript as submitted and take responsibility 6. Levitan EB, Liu S, Stampfer MJ, et al. HbA1c impaired fasting glucose, and impaired for the content of the article. measured in stored erythrocytes and mor- glucose tolerance: the Australian Diabetes, Parts of this study were presented in abstract tality rate among middle-aged and older Obesity, and Lifestyle Study (AusDiab). form at the 79th Congress of the European women. Diabetologia 2008;51:267–275 Circulation 2007;116:151–157 Atherosclerosis Society, Gothenburg, Sweden, 7. Sarwar N, Aspelund T, Eiriksdottir G, 14. Meigs JB, Nathan DM, D’Agostino RB Sr, 26–29 June 2011. et al. Markers of dysglycaemia and risk Wilson PW; Framingham Offspring Study. The authors are grateful to the participants of coronary heart disease in people with- Fasting and postchallenge glycemia and of the LURIC study; without their collabora- out diabetes: Reykjavik prospective study cardiovascular disease risk: the Framing- tion, this article would not have been written. and systematic review. PLoS Med 2010;7: ham Offspring Study. Diabetes Care 2002; The authors also thank the LURIC study team e1000278 25:1845–1850 either temporarily or permanently involved in 8. McNeely MJ, McClelland RL, Bild DE, et al. 15. Selvin E, Crainiceanu CM, Brancati FL, patient recruitment and sample and data han- The association between A1C and sub- Coresh J. Short-term variability in meas- dling, the laboratory staff at the Ludwigshafen clinical cardiovascular disease: the multi- ures of glycemia and implications for the General Hospital, and the Universities of ethnic study of atherosclerosis. Diabetes classification of diabetes. Arch Intern Med Freiburg and Ulm. Care 2009;32:1727–1733 2007;167:1545–1551

care.diabetesjournals.org DIABETES CARE 3