This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Transcript
Prognostic Value of Admission Glycosylated Hemoglobinand Glucose in Nondiabetic Patients With
Jorik R. Timmer, MD, PhD*; Miriam Hoekstra, MD*; Maarten W.N. Nijsten, MD, PhD;Iwan C.C. van der Horst, MD, PhD; Jan Paul Ottervanger, MD, PhD;
Robbert J. Slingerland, PhD; Jan-Henk E. Dambrink, MD, PhD; Henk J.G. Bilo, MD, PhD;Felix Zijlstra, MD, PhD; Arnoud W.J. van ’t Hof, MD, PhD
Background—In nondiabetic patients with ST-segment–elevation myocardial infarction, acute hyperglycemia is associ-ated with adverse outcome. Whether this association is due merely to hyperglycemia as an acute stress response orwhether longer-term glycometabolic derangements are also involved is uncertain. It was our aim to determine theassociation between both acute and chronic hyperglycemia (hemoglobin A1c [HbA1c]) and outcome in nondiabeticpatients with ST-segment–elevation myocardial infarction.
Methods and Results—This observational study included consecutive patients (n�4176) without known diabetes mellitusadmitted with ST-segment–elevation myocardial infarction. All patients were treated with primary percutaneousintervention. Both glucose and HbA1c were measured on admission. Main outcome measure was total long-termmortality; secondary outcome measures were 1-year mortality and enzymatic infarct size. One-year mortality was 4.7%,and mortality after total follow-up (3.3�1.5 years) was 10%. Both elevated HbA1c levels (P�0.001) and elevatedadmission glucose (P�0.001) were associated with 1-year and long-term mortality. After exclusion of early mortality(within 30 days), HbA1c remained associated with long-term mortality (P�0.001), whereas glucose lost significance(P�0.09). Elevated glucose, but not elevated HbA1c, was associated with larger infarct size. After multivariate analysis,HbA1c (hazard ratio, 1.2 per interquartile range; P�0.01), but not glucose, was independently associated with long-termmortality.
Conclusions—In nondiabetic patients with ST-segment–elevation myocardial infarction, both elevated admission glucoseand HbA1c levels were associated with adverse outcome. Both of these parameters reflect different patient populations,and their association with outcome is probably due to different mechanisms. Measurement of both parameters enablesidentification of these high-risk groups for aggressive secondary risk prevention. (Circulation. 2011;124:00-00.)
Prognosis after myocardial infarction in patients withdiabetes mellitus is worse compared with patients with-
out diabetes mellitus, even in the setting of optimal reperfu-sion strategy involving primary percutaneous intervention(PCI).1 Glycosylated hemoglobin (HbA1c) is an establishedmarker of long-term glycemic control in patients with diabe-tes mellitus, and elevated HbA1c levels in such patients areassociated with an increased risk for future microvascular andmacrovascular disease.2 Moreover, a recent report found thatelevated HbA1c levels are also predictive for cardiovasculardisease and mortality in patients without diabetes mellitus,
regardless of fasting glucose levels,3 indicating that long-termglycometabolic derangement in the subdiabetic range alsoposes a risk for cardiovascular disease.
Clinical Perspective on p ●●●
Acute glycometabolic derangement in nondiabetic patientswith myocardial infarction has already been proven to be apowerful predictor of prognosis.4–7 However, until now, dataon the predictive value of HbA1c levels, reflecting long-termglycometabolic control, in nondiabetic patients with myocar-dial infarction are limited.8–10 The aim of the present study
Received August 25, 2010; accepted May 16, 2011.From the Departments of Cardiology (J.R.T., J.P.O., J.-H.E.D., A.W.J.v.H.), Clinical Chemistry (R.J.S.), and Internal Medicine (H.J.G.B.), Isala
Klinieken, Zwolle, the Netherlands, and Departments of Anesthesiology (M.H.), Cardiology (M.H., I.C.C.v.d.H., F.Z.), Critical Care (M.W.N.N.), andInternal Medicine (H.J.G.B.), University Medical Center Groningen, University of Groningen, Groningen, the Netherlands.
Drs Timmer and Hoekstra contributed equally to this article.Correspondence to Arnoud W.J. van ’t Hof, MD, PhD, Department of Cardiology, Isala Klinieken, Groot Wezenland 20, 8011 JW Zwolle, Netherlands.
was to assess the prognostic impact of both admission HbA1c
and glucose levels in a large population of patients withoutknown diabetes mellitus who were treated with PCI forST-segment–elevation myocardial infarction (STEMI).
MethodsWe performed an observational study including all consecutivepatients admitted with ST-elevation myocardial infarction to 2 largehospitals (Isala Klinieken, Zwolle, and the University MedicalCenter Groningen, Groningen) in the Netherlands. The inclusionperiod was January 2004 to January 2009 for the hospital in Zwolleand January 2005 to April 2009 for the hospital in Groningen.During these time frames, HbA1c and admission glucose wereroutinely measured on admission in all STEMI patients.
ST-segment elevation myocardial infarction was defined as com-plaints of chest pain with ECG signs compatible with acute myocar-dial infarction (ST-segment elevation �2 mm in precordial leads and�1 mm in limb leads).11 All patients were directly transported to thecatheterization laboratory on arrival, and acute coronary angiographywas performed with subsequent PCI when indicated as part of theroutine treatment for all STEMI patients in these institutions. Theinterventional strategy was at the operator’s discretion. All patientswere pretreated with aspirin, heparin, and clopidogrel during trans-portation to the hospital, or these drugs were administered at theemergency ward.12
Data CollectionPatient characteristics were recorded on admission with either caserecord forms or a computer-based database. Ischemic time wasdefined as the time between symptom onset and first ballooninflation. Thrombolysis in Myocardial Infarction (TIMI) flow wasscored according to the TIMI flow grading system before and afterPCI.13 Myocardial blush grade was defined as previously de-scribed.14 Successful PCI was defined as TIMI grade 3 flow withmyocardial blush grade 2 to 3 after PCI. Myocardial infarct size wasmeasured by peak creatinine kinase level in the first 24 hours afteradmission. Diabetes mellitus was defined as known diabetes mellituson admission, which was treated with diet, oral glucose-loweringmedication, and/or insulin. Clinical follow-up was performed bytelephone contact (with either the general practitioner or the patient)or through coupling of municipal mortality records. Follow-up wasperformed by independent research nurses not involved in patienttreatment. The HbA1c levels were measured on the Primus Ultra 2affinity chromatography-HPLC (Primus Diagnostics, Kansas City,MO) in Zwolle with a within-run coefficient of variation of �0.5%and on a Roche COBAS Integra 800 closed-tube system in Groning-en. Both devices report the same reference normal values of 4.0% to6.0% in nondiabetics. Glucose levels were measured with a Modulardevice (Roche Diagnostics) in Zwolle and with a Radiometer ABL700/800 series analyzer (Radiometer Copenhagen) in full-bloodarterial samples or in sodium fluoride–containing tubes with theRoche Modular analyzer in Groningen. During the study period,reference values did not change, and yearly numeric quality controldata revealed that the coefficient of variation remained �2% duringthis time period. Both glucose measurements and HbA1c measure-ments were compared between the 2 centers. For optimal analysis,HbA1c levels were transformed linearly to match those of Zwolle.Measurements from the Zwolle hospital were used as the referencebecause the central laboratory of this center has extensive experiencewith the glucose and HbA1c assays and because this center contrib-uted the most patients. The HbA1c values from the UniversityMedical Center Groningen were corrected with a factor of 0.95 foroptimal matching with the patients from Zwolle. Glucose distribu-tions were similar between the 2 centers, and no adjustment wasnecessary.
To maintain a uniform patient population with genuine STEMI,specific inclusion and exclusion criteria were applied. To avoidinclusion of patients with a false diagnosis of STEMI (eg, owing topericarditis), only patients in whom a PCI was performed in the acutesetting were included. Patients who presented after an out-of-hospital
cardiac arrest were also excluded because prognosis in these patientsis driven primarily by neurological outcome. There were no exclu-sion criteria with regard to age, sex, ischemic time, cardiac history,or renal function.
Statistical AnalysisFor the analysis, patients with known diabetes mellitus were ex-cluded. The primary end point was long-term mortality (maximumfollow-up available per patient). Secondary end points were 1-yearmortality and enzymatic infarct size. Patient groups were createdaccording to quartiles of admission HbA1c and glucose, referred to asinterquartile range (IQR) 1 to 4. Continuous data were summarizedand are given as median values with corresponding IQR or as meanvalues with corresponding SD, and dichotomous data are given ascounts and percentages. Mortality data were compared by use ofeither �2 test (30-day and 1-year mortality) or log-rank analysis(long-term mortality) for comparison of Kaplan–Meier actuarialsurvival curves. Means between groups were compared by use ofindependent-samples t tests (ANOVA polynomial linear term) orMann–Whitney U tests (Kruskal-Wallis) when appropriate.
Kaplan–Meier curves were constructed for overall mortality usingthe log-rank statistic for comparisons between groups. In multivar-iate analysis (Cox regression using backward stepwise variableselection methods), the association between HbA1c, glucose andoutcome (long-term mortality) was adjusted for age, sex, and allpredictors of mortality (prior coronary artery disease, hypertension,active smoking, renal function, systolic blood pressure on admission,heart rate on admission, ischemic time, multivessel coronary arterydisease, anterior infarction, TIMI flow before PCI, TIMI flow afterPCI). To investigate the effect of glucose and HbA1c on early andlate mortality, a secondary univariate landmark analysis was per-formed excluding mortality within 30 days and resetting follow-uptimes after this period.
All statistical tests were performed with SPSS 12.0. A value ofP�0.05 was considered statistically significant.
ResultsFrom January 2004 until April 2009, a total of 5373 patientswere included: 3369 in Zwolle and 2004 in Groningen. Asdefined by the inclusion criteria, all patients were treated withprimary PCI. A total of 598 patients (11%) had diabetesmellitus on admission. Diabetic status was missing for 77patients (1.4%), and these patients were excluded from themain analysis. Of the remaining 4698 patients included in thepresent analysis, HbA1c on admission was not available in522 patients (11%). Final analysis was therefore performedon 4176 patients.
Mean age (62�13 versus 63�13 years; P�0.17) anddistribution of sex (male, 74% in both centers; P�0.82) andmedian HbA1c levels (5.60 [IQR, 5.40 to 5.80] versus 5.54[IQR, 2.26 to 5.92]; P�0.75) were comparable betweenpatients included from Zwolle and Groningen. Other baselineand angiographic characteristics were also comparable be-tween Zwolle and Groningen, except the presence of multi-vessel disease (45.8% versus 54.9%; P�0.001). For the totalgroup, the mean follow-up period was 3.3�1.5 years. Thirty-day follow-up was complete in 99.9% of the patients; 1-yearfollow-up was complete in 99.6% of the patients. One-yearmortality was 4.7% and long-term mortality was 10%.
Patients were divided on the basis of admission HbA1c
quartiles (IQR 1, �5.35%; IQR 2, 5.36% to 5.54%; IQR 3,5.55% to 5.80%; and IQR 4, �5.81%). Baseline and angio-graphic characteristics are shown in Table 1. Patients withhigher HbA1c levels were older, were more often female, andmore often had a prior history of coronary artery disease.
There was a strong correlation between admission HbA1c andadmission glucose level (P�0.001). Clinical outcome (30-day mortality, 1-year mortality, and infarct size) is displayedin Table 2. With increasing HbA1c levels, there was anincrease in the prevalence of multivessel disease, and therewas a modest increase in PCI failure in the upper quartile of
HbA1c. One-year mortality was significantly higher withincreasing HbA1c levels (P�0.001). Infarct size, as measuredby peak creatinine kinase, was not correlated with HbA1c
levels.Patients were also divided according to admission glucose
IQR 3, 8.2 to 9.5 mmol/L; and IQR 4, �9.6 mmol/L) (formg/dL, multiply by 18). Baseline and angiographic charac-teristics are shown in Table 3. Higher admission glucose wasassociated with more frequent presence of multivessel dis-ease, less frequent TIMI 3 flow on admission, and a lowerrate of successful PCI. There was a clear association betweenthe use of an intra-aortic balloon pump and high admissionglucose (P�0.001). Clinical outcome is shown in Table 2.Both 30-day mortality and long-term mortality were signifi-cantly associated with higher glucose levels (P�0.001). A Ushape was present with regard to admission glucose andmortality in which patients with low admission glucose (IQR1, �6.9 mmol/L) had a slightly higher mortality than patientswith normal admission glucose (IQR 2, 7.0 to 8.1 mmol/L).There was a significant positive correlation between admis-sion glucose and infarct size, measured by peak creatininekinase level (P�0.001).
One-year mortality in patients with known diabetes melli-tus (n�598) was 11.0% compared with 4.6% in the patients
without known diabetes mellitus on admission (n�4176;P�0.001). For patients without known diabetes mellitus onadmission, clinical outcome is presented in Table 2 stratifiedaccording to admission HbA1c and glucose IQR. Survivalcurves for admission HbA1c and glucose IQR for patientswithout known diabetes mellitus on admission are presentedin Figures 1 and 2, respectively. Elevated HbA1c levels weregenerally associated with increased long-term mortality re-gardless of admission glucose level, although the differencein patients with admission glucose levels in the highestquartile was not statistically significant (Figure 3). To inves-tigate the effect of glucose and HbA1c on early versus latemortality, we also performed an analysis after excludingpatients who died within the first 30 days. Glucose was nolonger associated with long-term mortality (P�0.09),whereas admission HbA1c remained significantly associatedwith long-term mortality (P�0.001).
After multivariate analysis correcting for baseline charac-teristics, hemodynamic parameters, and angiographic find-
Table 3. Baseline Characteristics of Nondiabetic Patients Based on Quartiles of Admission Glucose
IQR 1(�6.9 mmol/L,�124 mg/dL;
n�1034)
IQR 2(7.0–8.1 mmol/L,125–145 mg/dL;
n�1074)
IQR 3(8.2–9.5 mmol/L,145–171 mg/dL;
n�992)
IQR 4(�9.6 mmol/L,�172 mg/dL;
n�1032) P
Patient demographics
Age, mean�SD, y 60�13 61�13 63�12 65�12 �0.001
Male sex, % 76 77 74 69 �0.001
Body mass index, kg/m2 26.2 (24.0–28.9) 26.2 (24.2–28.6) 26.2 (24.3–28.4) 26.4 (24.5–29.2) NS
ings, HbA1c (hazard ratio per IQR, 1.2; 95% confidenceinterval, 1.0 to 1.3), but not admission glucose, was signifi-cantly associated with long-term mortality. Significant pre-dictors are presented in Table 4.
DiscussionOur study shows that in STEMI patients without knowndiabetes mellitus, both short- and long-term abnormalities inglucose control are associated with long-term mortality. Bothparameters reflect different patient populations, and theirassociation with outcome is probably due to different mech-anisms. Measurement of HbA1c levels in nondiabetic patientsmay improve risk assessment in patients presenting withacute STEMI.
Although acute hyperglycemia on admission and duringhospital stay has clearly been associated with adverse out-come in patients with acute myocardial infarction,5,15,16 theprognostic value of admission HbA1c levels in this popu-
lation has been less well established.8 –10 Our study showsthat admission HbA1c levels are associated with highermortality in a nondiabetic STEMI population treated withprimary PCI.
Several factors may play a role in the demonstratedassociation between HbA1c levels and adverse outcome.Increasing HbA1c levels were clearly associated with adversebaseline characteristics such as a higher cardiovascular riskprofile, explaining part of the increase in long-term mortality.In addition, it is conceivable that part of the associationbetween long-term abnormalities in glucose control andoutcome is due to the same complex mechanisms responsiblefor the adverse association between overt diabetes mellitusand cardiovascular outcome. Indeed, it has been well estab-lished that the excess risk for developing coronary arterydisease is not limited to patients with diabetes mellitus butalso is present in impaired fasting glucose, impaired glucosetolerance, and other states of insulin resistance.17–20 Our
Figure 1. Unadjusted Kaplan–Meier curves show-ing survival based on admission hemoglobin A1c(HbA1c) quartile in patients without diabetes melli-tus. IQR indicates interquartile range.
Figure 2. Unadjusted Kaplan–Meier curves show-ing survival based on admission glucose quartile inpatients without diabetes mellitus. IQR indicatesinterquartile range; PCI, percutaneous coronaryintervention.
Timmer et al Glycometabolic Derangements in Myocardial Infarction 5
findings indicate that these factors continue to play a negativerole after cardiovascular disease has become clinically overt.
Because the number of patients with long-term abnormal-ities in glucose control and subsequent cardiovascular se-quelae is likely to increase in the future decades, moretailored therapy should be investigated in this patient popu-lation. The European guidelines on diabetes mellitus, predi-abetes, and cardiovascular disease recommend that people athigh risk for type 2 diabetes mellitus should receive lifestylecounseling and, if needed, pharmacological therapy to reducetheir risk of developing overt hyperglycemia and type 2diabetes mellitus but especially to prevent or slow thedevelopment of cardiovascular disease.21,22 This approachcould also be encouraged in our patient population, and itmay alter prognosis, although the benefits with regard toslowing the progression to diabetes mellitus still have to beelucidated. However, it is known that the overall increase incardiovascular risk in patients with diabetes mellitus ormilder abnormalities in glucose levels is not explained byabnormalities in glucose or HbA1c alone, which is an impor-tant consideration in designing prevention efforts.
Specific strategies targeting the acute glucose abnormali-ties in STEMI may be beneficial in theory, but results of acute
interventions in glucose metabolism in patients with acutecoronary syndromes have proved disappointing.23–25 Moreconcise ideas regarding therapeutic implications and optionshave yet to evolve.
Hyperglycemia in STEMI patients was strongly associatedwith increased mortality. Although there is a clear correlationbetween admission glucose and HbA1c levels, they appear torepresent related but different phenomena. Patients withelevated glucose levels have larger myocardial infarctionsand less frequently have open infarct-related vessels on theinitial angiogram. They also need hemodynamic support of anintra-aortic balloon pump more often, probably reflectingsevere hemodynamic stress caused by pump failure. Indeed,after correction for hemodynamic parameters such as bloodpressure, heart rate on admission, and angiographic findings,glucose was no longer independently associated with long-term mortality.
In comparison, patients with elevated HbA1c levels partic-ularly had high-risk baseline characteristics such as a higherprevalence of prior cardiovascular disease and a higherprevalence of renal dysfunction. In these patients, there wasno increase in infarct size, nor did they need more mechanicalsupport of an intra-aortic balloon pump. So, it appears that themechanisms by which both glucose and HbA1c are linked tooutcome are distinct and may even be partially independentfrom each other. Indeed, in our study, glucose was particu-larly associated with mortality within 30 days. When patientswho died within 30 days were excluded, glucose lost itsassociation with mortality, whereas HbA1c remained a strongpredictor of future mortality. So, in contrast to HbA1c, thenegative impact of elevated admission glucose on prognosisis particularly reflected by early mortality. This probablyreflects the acute stress of hemodynamically unstable patientswith higher glucose levels compared with the more generalincrease in cardiovascular risk associated with higher HbA1c
levels.5
Recently, a prospective cohort study showed that in anondiabetic general population, an elevated HbA1c level is arisk factor for the development of cardiovascular eventsindependently of fasting glucose.3 Our data suggest thatHbA1c may also be used to assess cardiovascular risk in a
Figure 3. Bar graph showing unadjusted Kaplan–Meier–estimated 3-year mortality stratified onadmission glucose quartile and according tohemoglobin A1c (HbA1c) level (median value) inpatients without diabetes mellitus. P value wascalculated with log-rank analysis. IQR indicatesinterquartile range.
Table 4. Predictors of Long-Term Mortality in NondiabeticPatients by Multivariate Analysis
HR 95% CI P
Age (per decade) 2.0 1.8–2.3 �0.001
Male sex 1.4 1.1–1.8 0.02
Heart rate on admission �100 bpm 2.4 1.7–3.3 �0.001
Systolic blood pressure �100 mm Hg 1.9 1.4–2.7 �0.001
Absence of TIMI grade 3 flow after PCI 1.9 1.4–2.6 �0.001
HbA1c per IQR 1.2 1.0–1.3 �0.010
HR indicates hazard ratio; CI, confidence interval; TIMI, Thrombolysis inMyocardial Infarction; and IQR, interquartile range. TIMI grade 3 flow before PCIwas included in the model as a nonsignificant variable. The following candidatevariables were eliminated with backward stepwise variable selection methods:previous coronary artery disease, hypertension, active smoking, ischemic time,multivessel disease, anterior infarction, and admission glucose level.
nondiabetic population after STEMI. Thus, the importance ofelevated HbA1c in nondiabetics can be generalized from anindicator of primary risk to secondary risk. Besides providingprognostic information, routine HbA1c measurement inSTEMI patients may help to identify patients with undetecteddiabetes mellitus or those at increased risk for developingdiabetes mellitus in the future.3
Study LimitationsThis was a retrospective study without cause-specific mortal-ity. Diabetes mellitus was defined as known diabetic status onadmission. It is well known that a number of STEMI patientshave undetected diabetes mellitus, and they were not ex-cluded in our study.15,26 Indeed, when the HbA1c cutoff valueof �6.5% as suggested by the American Diabetes Associa-tion was used, �5% of our population could readily bediagnosed with diabetes mellitus on admission (undiagnoseddiabetes mellitus), and they were not excluded from the mainanalysis.27 However, because there was a stepwise increase inmortality with increasing HbA1c levels starting from thelowest to the highest quartile, we believe that our findings arenot solely attributable to the number of patients with unde-tected diabetes mellitus. It is conceivable that some subjectswithin the higher HbA1c quartiles may have progressed toovert diabetes mellitus within the follow-up period, whichwould also adversely affect long-term prognosis. Because wehad no data on the occurrence of this diagnosis or the start ofglucose-lowering medication during the follow-up period, themagnitude of this phenomenon was unknown in our study.Another limitation might be that although admission glucosewill be responsive to the acute stress associated with anSTEMI, many other factors may contribute to the variabilityof nonfasting glucose levels.
ConclusionsBoth elevated admission HbA1c and glucose were associatedwith an adverse prognosis in nondiabetic patients with ST-elevation myocardial infarction treated with primary PCI. TheHbA1c and glucose levels reflect different patient popula-tions, and their association with outcome is probably due todifferent mechanisms. High admission glucose is associatedwith a more hemodynamically unstable patient group with alarger infarct size and high early mortality. Elevated HbA1c isassociated with more adverse baseline characteristics and amore gradual higher mortality over time. Close follow-up ofthese patients seems warranted. More research is needed tobetter describe and understand these findings, and, moreimportant, to assess and develop feasible treatment options.
AcknowledgmentsVera Derks is credited for the submission process.
DisclosuresNone.
References1. Timmer JR, Ottervanger JP, de Boer MJ, Boersma E, Grines CL,
Westerhout CM, Simes RJ, Granger CB, Zijlstra F. Primary percutaneouscoronary intervention compared with fibrinolysis for myocardial
infarction in diabetes mellitus: results from the Primary Coronary Angio-plasty vs Thrombolysis-2 trial. Arch Intern Med. 2007;167:1353–1359.
2. Stolar M. Glycemic control and complications in type 2 diabetes mellitus.Am J Med. 2010;123:S3–S11.
3. Selvin E, Steffes MW, Zhu H, Matsushita K, Wagenknecht L, Pankow J,Coresh J, Brancati FL. Glycated hemoglobin, diabetes, and cardiovascu-lar risk in nondiabetic adults. N Engl J Med. 2010;362:800–811.
4. Goyal A, Mehta SR, Diaz R, Gerstein HC, Afzal R, Xavier D, Liu L, PaisP, Yusuf S. Differential clinical outcomes associated with hypoglycemiaand hyperglycemia in acute myocardial infarction. Circulation. 2009;120:2429–2437.
5. Capes SE, Hunt D, Malmberg K, Gerstein HC. Stress hyperglycaemia andincreased risk of death after myocardial infarction in patients with andwithout diabetes: a systematic overview. Lancet. 2000;355:773–778.
6. Kosiborod M, Rathore SS, Inzucchi SE, Masoudi FA, Wang Y, HavranekEP, Krumholz HM. Admission glucose and mortality in elderly patientshospitalized with acute myocardial infarction: implications for patientswith and without recognized diabetes. Circulation. 2005;111:3078–3086.
7. Kosiborod M. Blood glucose and its prognostic implications in patientshospitalised with acute myocardial infarction. Diab Vasc Dis Res. 2008;5:269–275.
8. Hadjadj S, Coisne D, Mauco G, Ragot S, Duengler F, Sosner P, Tor-remocha F, Herpin D, Marechaud R. Prognostic value of admissionplasma glucose and HbA in acute myocardial infarction. Diabet Med.2004;21:305–310.
9. Cakmak M, Cakmak N, Cetemen S, Tanriverdi H, Enc Y, Teskin O, KilicID. The value of admission glycosylated hemoglobin level in patientswith acute myocardial infarction. Can J Cardiol. 2008;24:375–378.
10. Tenerz A, Nilsson G, Forberg R, Ohrvik J, Malmberg K, Berne C,Leppert J. Basal glucometabolic status has an impact on long-termprognosis following an acute myocardial infarction in non-diabeticpatients. J Intern Med. 2003;254:494–503.
11. Thygesen K, Alpert JS, White HD, Jaffe AS, Apple FS, Galvani M, KatusHA, Newby LK, Ravkilde J, Chaitman B, Clemmensen PM, Dellborg M,Hod H, Porela P, Underwood R, Bax JJ, Beller GA, Bonow R, Van derWall EE, Bassand JP, Wijns W, Ferguson TB, Steg PG, Uretsky BF,Williams DO, Armstrong PW, Antman EM, Fox KA, Hamm CW, OhmanEM, Simoons ML, Poole-Wilson PA, Gurfinkel EP, Lopez-Sendon JL,Pais P, Mendis S, Zhu JR, Wallentin LC, Fernandez-Aviles F, Fox KM,Parkhomenko AN, Priori SG, Tendera M, Voipio-Pulkki LM, VahanianA, Camm AJ, De Caterina R, Dean V, Dickstein K, Filippatos G, Funck-Brentano C, Hellemans I, Kristensen SD, McGregor K, Sechtem U, SilberS, Tendera M, Widimsky P, Zamorano JL, Morais J, Brener S, HarringtonR, Morrow D, Lim M, Martinez-Rios MA, Steinhubl S, Levine GN,Gibler WB, Goff D, Tubaro M, Dudek D, Al-Attar N. Universal defi-nition of myocardial infarction. Circulation. 2007;116:2634–2653.
12. Van de Werf F, Bax J, Betriu A, Blomstrom-Lundqvist C, Crea F, FalkV, Filippatos G, Fox K, Huber K, Kastrati A, Rosengren A, Steg PG,Tubaro M, Verheugt F, Weidinger F, Weis M. Management of acutemyocardial infarction in patients presenting with persistent ST-segmentelevation: the Task Force on the Management of ST-Segment ElevationAcute Myocardial Infarction of the European Society of Cardiology. EurHeart J. 2008;29:2909–2945.
13. The Thrombolysis in Myocardial Infarction (TIMI) trial: phase I findings:TIMI Study Group. N Engl J Med. 1985;312:932–936.
14. Henriques JP, Zijlstra F, van ’t Hof AWJ, de Boer MJ, Dambrink JHE,Gosselink ATM, Hoorntje JCA, Suryapranata H. Angiographicassessment of reperfusion in acute myocardial infarction by myocardialblush grade. Circulation. 2003;107:2115–2119.
15. Norhammar A, Tenerz A, Nilsson G, Hamsten A, Efendic S, Ryden L,Malmberg K. Glucose metabolism in patients with acute myocardialinfarction and no previous diagnosis of diabetes mellitus: a prospectivestudy. Lancet. 2002;359:2140–2144.
16. Timmer JR, van der Horst I, Ottervanger JP, Henriques JP, Hoorntje JCA,de Boer MJ, Suryapranata H, Zijlstra F. Prognostic value of admissionglucose in non-diabetic patients with myocardial infarction. Am Heart J.2004;148:399–404.
17. DECODE Study Group. Is the current definition for diabetes relevant tomortality risk from all causes and cardiovascular and noncardiovasculardiseases? Diabetes Care. 2003;26:688–696.
18. Coutinho M, Gerstein HC, Wang Y, Yusuf S. The relationship betweenglucose and incident cardiovascular events: a metaregression analysis ofpublished data from 20 studies of 95,783 individuals followed for 124years. Diabetes Care. 1999;22:233–240.
Timmer et al Glycometabolic Derangements in Myocardial Infarction 7
20. Khaw KT, Wareham N, Bingham S, Luben R, Welch A, Day N. Asso-ciation of hemoglobin A1c with cardiovascular disease and mortality inadults: the European Prospective Investigation Into Cancer in Norfolk.Ann Intern Med. 2004;141:413–420.
21. Ryden L, Standl E, Bartnik M, Van den Berghe G, Betteridge J, de BoerMJ, Cosentino F, Jonsson B, Laakso M, Malmberg K, Priori S, OstergrenJ, Tuomilehto J, Thrainsdottir I, Vanhorebeek I, Stramba-Badiale M,Lindgren P, Qiao Q, Priori SG, Blanc JJ, Budaj A, Camm J, Dean V,Deckers J, Dickstein K, Lekakis J, McGregor K, Metra M, Morais J,Osterspey A, Tamargo J, Zamorano JL, Deckers JW, Bertrand M, Char-bonnel B, Erdmann E, Ferrannini E, Flyvbjerg A, Gohlke H, Juanatey JR,Graham I, Monteiro PF, Parhofer K, Pyorala K, Raz I, Schernthaner G,Volpe M, Wood D. Guidelines on diabetes, pre-diabetes, and cardiovas-cular diseases: executive summary: the Task Force on Diabetes andCardiovascular Diseases of the European Society of Cardiology (ESC)and of the European Association for the Study of Diabetes (EASD). EurHeart J. 2007;28:88–136.
22. Orchard TJ, Temprosa M, Goldberg R, Haffner S, Ratner R, MarcovinaS, Fowler S. The effect of metformin and intensive lifestyle intervention
on the metabolic syndrome: the Diabetes Prevention Program randomizedtrial. Ann Intern Med. 2005;142:611–619.
23. Timmer JR, Svilaas T, Ottervanger JP, Henriques JP, Dambrink JHE, vanden Broek SA, van der Horst I, Zijlstra F. Glucose-insulin-potassiuminfusion in patients with acute myocardial infarction without signs ofheart failure: the Glucose-Insulin-Potassium Study (GIPS)-II. J Am CollCardiol. 2006;47:1730–1731.
24. Malmberg K, Ryden L, Wedel H, Birkeland K, Bootsma A, Dickstein K,Efendic S, Fisher M, Hamsten A, Herlitz J, Hildebrandt P, MacLeod K,Laakso M, Torp-Pedersen C, Waldenstrom A. Intense metabolic controlby means of insulin in patients with diabetes mellitus and acute myo-cardial infarction (DIGAMI 2): effects on mortality and morbidity. EurHeart J. 2005;26:650–661.
25. Diaz R, Goyal A, Mehta SR, Afzal R, Xavier D, Pais P, Chrolavicius S,Zhu J, Kazmi K, Liu L, Budaj A, Zubaid M, Avezum A, Ruda M, YusufS. Glucose-insulin-potassium therapy in patients with ST-segment ele-vation myocardial infarction. JAMA. 2007;298:2399–2405.
26. Knudsen EC, Seljeflot I, Abdelnoor M, Eritsland J, Mangschau A,Arnesen H, Andersen GO. Abnormal glucose regulation in patients withacute ST-elevation myocardial infarction: a cohort study on 224 patients.Cardiovasc Diabetol. 2009;8:6.
27. Diagnosis and classification of diabetes mellitus. Diabetes Care. 2010;33(suppl 1):S62–S69.
CLINICAL PERSPECTIVEMeasurement of admission glucose and hemoglobin A1c (HbA1c) in acute myocardial infarction may identify patients withdisturbed glucose metabolism and an increased risk for adverse outcome. Although HbA1c and glucose are related, theycan differentiate between mechanisms of adverse outcome. Admission glucose is related to increased hemodynamic stress,whereas HbA1c identifies patients with higher long-term cardiovascular risk, possibly by abnormal long-term glucoselevels. Early identification of these patient groups enables the initiation of specific intervention strategies and may help usdevelop strategies to improve prognosis in these high-risk patient groups. This is of particular importance because thereis a global increase in the number of patients suffering from cardiovascular disease with underlying insulin resistance,prediabetes, and overt diabetes mellitus. Both glucose and HbA1c should be measured in patients admitted withST-segment–elevation myocardial infarction.
Arnoud W.J. van 't HofOttervanger, Robbert J. Slingerland, Jan-Henk E. Dambrink, Henk J.G. Bilo, Felix Zijlstra and Jorik R. Timmer, Miriam Hoekstra, Maarten W.N. Nijsten, Iwan C.C. van der Horst, Jan Paul
Coronary InterventionElevation Myocardial Infarction Treated With Percutaneous−Patients With ST-Segment
Prognostic Value of Admission Glycosylated Hemoglobin and Glucose in Nondiabetic
is online at: Circulation Information about subscribing to Subscriptions:
http://www.lww.com/reprints Information about reprints can be found online at: Reprints:
document. Permissions and Rights Question and Answer this process is available in the
click Request Permissions in the middle column of the Web page under Services. Further information aboutOffice. Once the online version of the published article for which permission is being requested is located,
can be obtained via RightsLink, a service of the Copyright Clearance Center, not the EditorialCirculationin Requests for permissions to reproduce figures, tables, or portions of articles originally publishedPermissions:
References Selvin E, Steffes MW, Zhu H, Matsushita K, Wagenknecht L, Pankow J, Coresh J, 1. Brancati FL. Glycated hemoglobin, diabetes, and cardiovascular risk in non-diabetic adults. N Engl J Med. 2010;362:800-811.Rydén L, Standl E, Bartnik M, Van den Berghe G, Betteridge J, de Boer MJ, 2. Cosentino F, Jönsson B, Laakso M, Malmberg K, Priori S, Ostergren J, Tuomilehto J, Thrainsdottir I, Vanhorebeek I, Stramba-Badiale M, Lindgren P, Qiao Q, Priori SG, Blanc JJ, Budaj A, Camm J, Dean V, Deckers J, Dickstein K, Lekakis J, McGregor K, Metra M, Morais J, Osterspey A, Tamargo J, Zamorano JL, Deckers JW, Bertrand M, Charbonnel B, Erdmann E, Ferrannini E, Flyvbjerg A, Gohlke H, Juanatey JR, Graham I, Monteiro PF, Parhofer K, Pyörälä K, Raz I, Schernthaner G, Volpe M, Wood D; Task Force on Diabetes and Cardiovascular Diseases of the European Society of Cardiology (ESC); European Association for the Study of Diabetes (EASD). Guidelines on diabetes, prediabetes, and cardiovascular disease: executive summary: the Task Force on Diabetes and Cardiovascular Diseases of the European Society of Cardiology (ESC) and of the European Association for the Study of Diabetes (EASD). Eur Heart J. 2007;28:88-136.
85
Prognostic Value of Admission Glycosylated Hemoglobinand Glucose in Nondiabetic Patients With
Jorik R. Timmer, MD, PhD*; Miriam Hoekstra, MD*; Maarten W.N. Nijsten, MD, PhD;Iwan C.C. van der Horst, MD, PhD; Jan Paul Ottervanger, MD, PhD;
Robbert J. Slingerland, PhD; Jan-Henk E. Dambrink, MD, PhD; Henk J.G. Bilo, MD, PhD;Felix Zijlstra, MD, PhD; Arnoud W.J. van ’t Hof, MD, PhD
Background—In nondiabetic patients with ST-segment–elevation myocardial infarction, acute hyperglycemia is associ-ated with adverse outcome. Whether this association is due merely to hyperglycemia as an acute stress response orwhether longer-term glycometabolic derangements are also involved is uncertain. It was our aim to determine theassociation between both acute and chronic hyperglycemia (hemoglobin A1c [HbA1c]) and outcome in nondiabeticpatients with ST-segment–elevation myocardial infarction.
Methods and Results—This observational study included consecutive patients (n�4176) without known diabetes mellitusadmitted with ST-segment–elevation myocardial infarction. All patients were treated with primary percutaneousintervention. Both glucose and HbA1c were measured on admission. Main outcome measure was total long-termmortality; secondary outcome measures were 1-year mortality and enzymatic infarct size. One-year mortality was 4.7%,and mortality after total follow-up (3.3�1.5 years) was 10%. Both elevated HbA1c levels (P�0.001) and elevated admissionglucose (P�0.001) were associated with 1-year and long-term mortality. After exclusion of early mortality (within 30 days),HbA1c remained associated with long-term mortality (P�0.001), whereas glucose lost significance (P�0.09). Elevatedglucose, but not elevated HbA1c, was associated with larger infarct size. After multivariate analysis, HbA1c (hazard ratio, 1.2per interquartile range; P�0.01), but not glucose, was independently associated with long-term mortality.
Conclusions—In nondiabetic patients with ST-segment–elevation myocardial infarction, both elevated admission glucoseand HbA1c levels were associated with adverse outcome. Both of these parameters reflect different patient populations,and their association with outcome is probably due to different mechanisms. Measurement of both parameters enablesidentification of these high-risk groups for aggressive secondary risk prevention. (Circulation. 2011;124:704-711.)
Prognosis after myocardial infarction in patients withdiabetes mellitus is worse compared with patients with-
out diabetes mellitus, even in the setting of optimal reperfu-sion strategy involving primary percutaneous intervention(PCI).1 Glycosylated hemoglobin (HbA1c) is an establishedmarker of long-term glycemic control in patients with diabe-tes mellitus, and elevated HbA1c levels in such patients areassociated with an increased risk for future microvascular andmacrovascular disease.2 Moreover, a recent report found thatelevated HbA1c levels are also predictive for cardiovasculardisease and mortality in patients without diabetes mellitus,
regardless of fasting glucose levels,3 indicating that long-termglycometabolic derangement in the subdiabetic range alsoposes a risk for cardiovascular disease.
Clinical Perspective on p 92
Acute glycometabolic derangement in nondiabetic patientswith myocardial infarction has already been proven to be apowerful predictor of prognosis.4–7 However, until now, dataon the predictive value of HbA1c levels, reflecting long-termglycometabolic control, in nondiabetic patients with myocar-
Continuing medical education (CME) credit is available for this article. Go to http://cme.ahajournals.org to take the quiz.Received August 25, 2010; accepted May 16, 2011.From the Departments of Cardiology (J.R.T., J.P.O., J.-H.E.D., A.W.J.v.H.), Clinical Chemistry (R.J.S.), and Internal Medicine (H.J.G.B.), Isala
Klinieken, Zwolle, the Netherlands, and Departments of Anesthesiology (M.H.), Cardiology (M.H., I.C.C.v.d.H., F.Z.), Critical Care (M.W.N.N.), andInternal Medicine (H.J.G.B.), University Medical Center Groningen, University of Groningen, Groningen, the Netherlands.
Drs Timmer and Hoekstra contributed equally to this article.Correspondence to Arnoud W.J. van ’t Hof, MD, PhD, Department of Cardiology, Isala Klinieken, Groot Wezenland 20, 8011 JW Zwolle, Netherlands.
Circulation is available at http://circ.ahajournals.org DOI: 10.1161/CIRCULATIONAHA.110.985911
704 by IMED Korea on November 13, 2011http://circ.ahajournals.org/Downloaded from
86
dial infarction are limited.8–10 The aim of the present studywas to assess the prognostic impact of both admission HbA1c
and glucose levels in a large population of patients withoutknown diabetes mellitus who were treated with PCI forST-segment–elevation myocardial infarction (STEMI).
MethodsWe performed an observational study including all consecutivepatients admitted with ST-elevation myocardial infarction to 2 largehospitals (Isala Klinieken, Zwolle, and the University MedicalCenter Groningen, Groningen) in the Netherlands. The inclusionperiod was January 2004 to January 2009 for the hospital in Zwolleand January 2005 to April 2009 for the hospital in Groningen.During these time frames, HbA1c and admission glucose wereroutinely measured on admission in all STEMI patients.
ST-segment elevation myocardial infarction was defined as com-plaints of chest pain with ECG signs compatible with acute myocar-dial infarction (ST-segment elevation �2 mm in precordial leads and�1 mm in limb leads).11 All patients were directly transported to thecatheterization laboratory on arrival, and acute coronary angiographywas performed with subsequent PCI when indicated as part of theroutine treatment for all STEMI patients in these institutions. Theinterventional strategy was at the operator’s discretion. All patientswere pretreated with aspirin, heparin, and clopidogrel during trans-portation to the hospital, or these drugs were administered at theemergency ward.12
Data CollectionPatient characteristics were recorded on admission with either caserecord forms or a computer-based database. Ischemic time wasdefined as the time between symptom onset and first ballooninflation. Thrombolysis in Myocardial Infarction (TIMI) flow wasscored according to the TIMI flow grading system before and afterPCI.13 Myocardial blush grade was defined as previously de-scribed.14 Successful PCI was defined as TIMI grade 3 flow withmyocardial blush grade 2 to 3 after PCI. Myocardial infarct size wasmeasured by peak creatinine kinase level in the first 24 hours afteradmission. Diabetes mellitus was defined as known diabetes mellituson admission, which was treated with diet, oral glucose-loweringmedication, and/or insulin. Clinical follow-up was performed bytelephone contact (with either the general practitioner or the patient)or through coupling of municipal mortality records. Follow-up wasperformed by independent research nurses not involved in patienttreatment. The HbA1c levels were measured on the Primus Ultra 2affinity chromatography-HPLC (Primus Diagnostics, Kansas City,MO) in Zwolle with a within-run coefficient of variation of �0.5%and on a Roche COBAS Integra 800 closed-tube system in Groning-en. Both devices report the same reference normal values of 4.0% to6.0% in nondiabetics. Glucose levels were measured with a Modulardevice (Roche Diagnostics) in Zwolle and with a Radiometer ABL700/800 series analyzer (Radiometer Copenhagen) in full-bloodarterial samples or in sodium fluoride–containing tubes with theRoche Modular analyzer in Groningen. During the study period,reference values did not change, and yearly numeric quality controldata revealed that the coefficient of variation remained �2% duringthis time period. Both glucose measurements and HbA1c measure-ments were compared between the 2 centers. For optimal analysis,HbA1c levels were transformed linearly to match those of Zwolle.Measurements from the Zwolle hospital were used as the referencebecause the central laboratory of this center has extensive experiencewith the glucose and HbA1c assays and because this center contrib-uted the most patients. The HbA1c values from the UniversityMedical Center Groningen were corrected with a factor of 0.95 foroptimal matching with the patients from Zwolle. Glucose distribu-tions were similar between the 2 centers, and no adjustment wasnecessary.
To maintain a uniform patient population with genuine STEMI,specific inclusion and exclusion criteria were applied. To avoidinclusion of patients with a false diagnosis of STEMI (eg, owing to
pericarditis), only patients in whom a PCI was performed in the acutesetting were included. Patients who presented after an out-of-hospitalcardiac arrest were also excluded because prognosis in these patientsis driven primarily by neurological outcome. There were no exclu-sion criteria with regard to age, sex, ischemic time, cardiac history,or renal function.
Statistical AnalysisFor the analysis, patients with known diabetes mellitus were ex-cluded. The primary end point was long-term mortality (maximumfollow-up available per patient). Secondary end points were 1-yearmortality and enzymatic infarct size. Patient groups were createdaccording to quartiles of admission HbA1c and glucose, referred to asinterquartile range (IQR) 1 to 4. Continuous data were summarizedand are given as median values with corresponding IQR or as meanvalues with corresponding SD, and dichotomous data are given ascounts and percentages. Mortality data were compared by use ofeither �2 test (30-day and 1-year mortality) or log-rank analysis(long-term mortality) for comparison of Kaplan–Meier actuarialsurvival curves. Means between groups were compared by use ofindependent-samples t tests (ANOVA polynomial linear term) orMann–Whitney U tests (Kruskal-Wallis) when appropriate.
Kaplan–Meier curves were constructed for overall mortality usingthe log-rank statistic for comparisons between groups. In multivar-iate analysis (Cox regression using backward stepwise variableselection methods), the association between HbA1c, glucose andoutcome (long-term mortality) was adjusted for age, sex, and allpredictors of mortality (prior coronary artery disease, hypertension,active smoking, renal function, systolic blood pressure on admission,heart rate on admission, ischemic time, multivessel coronary arterydisease, anterior infarction, TIMI flow before PCI, TIMI flow afterPCI). To investigate the effect of glucose and HbA1c on early andlate mortality, a secondary univariate landmark analysis was per-formed excluding mortality within 30 days and resetting follow-uptimes after this period.
All statistical tests were performed with SPSS 12.0. A value ofP�0.05 was considered statistically significant.
ResultsFrom January 2004 until April 2009, a total of 5373 patientswere included: 3369 in Zwolle and 2004 in Groningen. Asdefined by the inclusion criteria, all patients were treated withprimary PCI. A total of 598 patients (11%) had diabetesmellitus on admission. Diabetic status was missing for 77patients (1.4%), and these patients were excluded from themain analysis. Of the remaining 4698 patients included in thepresent analysis, HbA1c on admission was not available in522 patients (11%). Final analysis was therefore performedon 4176 patients.
Mean age (62�13 versus 63�13 years; P�0.17) anddistribution of sex (male, 74% in both centers; P�0.82) andmedian HbA1c levels (5.60 [IQR, 5.40 to 5.80] versus 5.54[IQR, 2.26 to 5.92]; P�0.75) were comparable betweenpatients included from Zwolle and Groningen. Other baselineand angiographic characteristics were also comparable be-tween Zwolle and Groningen, except the presence of multi-vessel disease (45.8% versus 54.9%; P�0.001). For the totalgroup, the mean follow-up period was 3.3�1.5 years. Thirty-day follow-up was complete in 99.9% of the patients; 1-yearfollow-up was complete in 99.6% of the patients. One-yearmortality was 4.7% and long-term mortality was 10%.
Patients were divided on the basis of admission HbA1c
quartiles (IQR 1, �5.35%; IQR 2, 5.36% to 5.54%; IQR 3,5.55% to 5.80%; and IQR 4, �5.81%). Baseline and angio-graphic characteristics are shown in Table 1. Patients with
Timmer et al Glycometabolic Derangements in Myocardial Infarction 705
by IMED Korea on November 13, 2011http://circ.ahajournals.org/Downloaded from
87
higher HbA1c levels were older, were more often female, andmore often had a prior history of coronary artery disease.There was a strong correlation between admission HbA1c andadmission glucose level (P�0.001). Clinical outcome (30-day mortality, 1-year mortality, and infarct size) is displayedin Table 2. With increasing HbA1c levels, there was an
increase in the prevalence of multivessel disease, and therewas a modest increase in PCI failure in the upper quartile ofHbA1c. One-year mortality was significantly higher withincreasing HbA1c levels (P�0.001). Infarct size, as measuredby peak creatinine kinase, was not correlated with HbA1c
levels.
Table 1. Baseline Characteristics of Nondiabetic Patients Based on Quartiles of AdmissionHemoglobin A1c Levels
IQR 1(�5.35%;n�1119)
IQR 2(5.36%–5.54%;
n�991)
IQR 3(5.55%–5.80%;
n�1042)
IQR 4(�5.81%;n�1024) P
Patient demographics
Age, mean�SD, y 59�12 62�12 63�13 65�12 �0.001
Male sex, % 77 77 73 69 �0.001
Body mass index, kg/m2 25.7 (23.7–28.1) 26.0 (24.1–28.4) 26.2 (24.2–28.7) 27.2 (24.8–30.1) �0.001
IQR, interquartile range; CK, creatinine kinase. Values are expressed as median (IQR) or group percentage unless otherwise specified.
706 Circulation August 9, 2011
by IMED Korea on November 13, 2011http://circ.ahajournals.org/Downloaded from
88
Patients were also divided according to admission glucosequartiles (IQR 1, �6.9 mmol/L; IQR 2, 7.0 to 8.1 mmol/L;IQR 3, 8.2 to 9.5 mmol/L; and IQR 4, �9.6 mmol/L) (formg/dL, multiply by 18). Baseline and angiographic charac-teristics are shown in Table 3. Higher admission glucose wasassociated with more frequent presence of multivessel dis-ease, less frequent TIMI 3 flow on admission, and a lowerrate of successful PCI. There was a clear association betweenthe use of an intra-aortic balloon pump and high admissionglucose (P�0.001). Clinical outcome is shown in Table 2.Both 30-day mortality and long-term mortality were signifi-cantly associated with higher glucose levels (P�0.001). A Ushape was present with regard to admission glucose andmortality in which patients with low admission glucose (IQR1, �6.9 mmol/L) had a slightly higher mortality than patientswith normal admission glucose (IQR 2, 7.0 to 8.1 mmol/L).There was a significant positive correlation between admis-sion glucose and infarct size, measured by peak creatininekinase level (P�0.001).
One-year mortality in patients with known diabetes melli-tus (n�598) was 11.0% compared with 4.6% in the patientswithout known diabetes mellitus on admission (n�4176;P�0.001). For patients without known diabetes mellitus onadmission, clinical outcome is presented in Table 2 stratifiedaccording to admission HbA1c and glucose IQR. Survivalcurves for admission HbA1c and glucose IQR for patientswithout known diabetes mellitus on admission are presentedin Figures 1 and 2, respectively. Elevated HbA1c levels weregenerally associated with increased long-term mortality re-gardless of admission glucose level, although the differencein patients with admission glucose levels in the highestquartile was not statistically significant (Figure 3). To inves-tigate the effect of glucose and HbA1c on early versus latemortality, we also performed an analysis after excludingpatients who died within the first 30 days. Glucose was nolonger associated with long-term mortality (P�0.09),whereas admission HbA1c remained significantly associatedwith long-term mortality (P�0.001).
Table 3. Baseline Characteristics of Nondiabetic Patients Based on Quartiles of Admission Glucose
IQR 1(�6.9 mmol/L,�124 mg/dL;
n�1034)
IQR 2(7.0–8.1 mmol/L,125–145 mg/dL;
n�1074)
IQR 3(8.2–9.5 mmol/L,145–171 mg/dL;
n�992)
IQR 4(�9.6 mmol/L,�172 mg/dL;
n�1032) P
Patient demographics
Age, mean�SD, y 60�13 61�13 63�12 65�12 �0.001
Male sex, % 76 77 74 69 �0.001
Body mass index, kg/m2 26.2 (24.0–28.9) 26.2 (24.2–28.6) 26.2 (24.3–28.4) 26.4 (24.5–29.2) NS
IQR indicates interquartile range; PCI, percutaneous coronary intervention; CABG, coronary artery bypass grafting; BP, bloodpressure; eGFR, estimated glomerular filtration rate; LAD, left anterior descending coronary artery; TIMI, Thrombolysis in MyocardialInfarction; and IABP, intra-aortic balloon pump. Values are expressed as median (IQR) or group percentage unless otherwise specified.
Timmer et al Glycometabolic Derangements in Myocardial Infarction 707
by IMED Korea on November 13, 2011http://circ.ahajournals.org/Downloaded from
89
After multivariate analysis correcting for baseline charac-teristics, hemodynamic parameters, and angiographic find-ings, HbA1c (hazard ratio per IQR, 1.2; 95% confidenceinterval, 1.0 to 1.3), but not admission glucose, was signifi-cantly associated with long-term mortality. Significant pre-dictors are presented in Table 4.
DiscussionOur study shows that in STEMI patients without knowndiabetes mellitus, both short- and long-term abnormalities inglucose control are associated with long-term mortality. Bothparameters reflect different patient populations, and theirassociation with outcome is probably due to different mech-anisms. Measurement of HbA1c levels in nondiabetic patientsmay improve risk assessment in patients presenting withacute STEMI.
Although acute hyperglycemia on admission and duringhospital stay has clearly been associated with adverse out-
come in patients with acute myocardial infarction,5,15,16 theprognostic value of admission HbA1c levels in this popu-lation has been less well established.8 –10 Our study showsthat admission HbA1c levels are associated with highermortality in a nondiabetic STEMI population treated withprimary PCI.
Several factors may play a role in the demonstratedassociation between HbA1c levels and adverse outcome.Increasing HbA1c levels were clearly associated with adversebaseline characteristics such as a higher cardiovascular riskprofile, explaining part of the increase in long-term mortality.In addition, it is conceivable that part of the associationbetween long-term abnormalities in glucose control andoutcome is due to the same complex mechanisms responsiblefor the adverse association between overt diabetes mellitusand cardiovascular outcome. Indeed, it has been well estab-lished that the excess risk for developing coronary arterydisease is not limited to patients with diabetes mellitus but
Figure 1. Unadjusted Kaplan–Meier curves show-ing survival based on admission hemoglobin A1c(HbA1c) quartile in patients without diabetes melli-tus. IQR indicates interquartile range.
Figure 2. Unadjusted Kaplan–Meier curves show-ing survival based on admission glucose quartile inpatients without diabetes mellitus. IQR indicatesinterquartile range; PCI, percutaneous coronaryintervention.
708 Circulation August 9, 2011
by IMED Korea on November 13, 2011http://circ.ahajournals.org/Downloaded from
90
also is present in impaired fasting glucose, impaired glucosetolerance, and other states of insulin resistance.17–20 Ourfindings indicate that these factors continue to play a negativerole after cardiovascular disease has become clinically overt.
Because the number of patients with long-term abnormal-ities in glucose control and subsequent cardiovascular se-quelae is likely to increase in the future decades, moretailored therapy should be investigated in this patient popu-lation. The European guidelines on diabetes mellitus, predi-abetes, and cardiovascular disease recommend that people athigh risk for type 2 diabetes mellitus should receive lifestylecounseling and, if needed, pharmacological therapy to reducetheir risk of developing overt hyperglycemia and type 2diabetes mellitus but especially to prevent or slow thedevelopment of cardiovascular disease.21,22 This approachcould also be encouraged in our patient population, and itmay alter prognosis, although the benefits with regard toslowing the progression to diabetes mellitus still have to beelucidated. However, it is known that the overall increase incardiovascular risk in patients with diabetes mellitus ormilder abnormalities in glucose levels is not explained byabnormalities in glucose or HbA1c alone, which is an impor-tant consideration in designing prevention efforts.
Specific strategies targeting the acute glucose abnormali-ties in STEMI may be beneficial in theory, but results of acuteinterventions in glucose metabolism in patients with acutecoronary syndromes have proved disappointing.23–25 Moreconcise ideas regarding therapeutic implications and optionshave yet to evolve.
Hyperglycemia in STEMI patients was strongly associatedwith increased mortality. Although there is a clear correlationbetween admission glucose and HbA1c levels, they appear torepresent related but different phenomena. Patients withelevated glucose levels have larger myocardial infarctionsand less frequently have open infarct-related vessels on theinitial angiogram. They also need hemodynamic support of anintra-aortic balloon pump more often, probably reflectingsevere hemodynamic stress caused by pump failure. Indeed,after correction for hemodynamic parameters such as bloodpressure, heart rate on admission, and angiographic findings,glucose was no longer independently associated with long-term mortality.
In comparison, patients with elevated HbA1c levels partic-ularly had high-risk baseline characteristics such as a higherprevalence of prior cardiovascular disease and a higherprevalence of renal dysfunction. In these patients, there wasno increase in infarct size, nor did they need more mechanicalsupport of an intra-aortic balloon pump. So, it appears that themechanisms by which both glucose and HbA1c are linked tooutcome are distinct and may even be partially independentfrom each other. Indeed, in our study, glucose was particu-larly associated with mortality within 30 days. When patientswho died within 30 days were excluded, glucose lost itsassociation with mortality, whereas HbA1c remained a strongpredictor of future mortality. So, in contrast to HbA1c, thenegative impact of elevated admission glucose on prognosisis particularly reflected by early mortality. This probablyreflects the acute stress of hemodynamically unstable patientswith higher glucose levels compared with the more generalincrease in cardiovascular risk associated with higher HbA1c
levels.5
Recently, a prospective cohort study showed that in anondiabetic general population, an elevated HbA1c level is arisk factor for the development of cardiovascular events
Figure 3. Bar graph showing unadjusted Kaplan–Meier–estimated 3-year mortality stratified onadmission glucose quartile and according tohemoglobin A1c (HbA1c) level (median value) inpatients without diabetes mellitus. P value wascalculated with log-rank analysis. IQR indicatesinterquartile range.
Table 4. Predictors of Long-Term Mortality in NondiabeticPatients by Multivariate Analysis
HR 95% CI P
Age (per decade) 2.0 1.8–2.3 �0.001
Male sex 1.4 1.1–1.8 0.02
Heart rate on admission �100 bpm 2.4 1.7–3.3 �0.001
Systolic blood pressure �100 mm Hg 1.9 1.4–2.7 �0.001
Absence of TIMI grade 3 flow after PCI 1.9 1.4–2.6 �0.001
HbA1c per IQR 1.2 1.0–1.3 �0.010
HR indicates hazard ratio; CI, confidence interval; TIMI, Thrombolysis inMyocardial Infarction; and IQR, interquartile range. TIMI grade 3 flow before PCIwas included in the model as a nonsignificant variable. The following candidatevariables were eliminated with backward stepwise variable selection methods:previous coronary artery disease, hypertension, active smoking, ischemic time,multivessel disease, anterior infarction, and admission glucose level.
Timmer et al Glycometabolic Derangements in Myocardial Infarction 709
by IMED Korea on November 13, 2011http://circ.ahajournals.org/Downloaded from
91
independently of fasting glucose.3 Our data suggest thatHbA1c may also be used to assess cardiovascular risk in anondiabetic population after STEMI. Thus, the importance ofelevated HbA1c in nondiabetics can be generalized from anindicator of primary risk to secondary risk. Besides providingprognostic information, routine HbA1c measurement inSTEMI patients may help to identify patients with undetecteddiabetes mellitus or those at increased risk for developingdiabetes mellitus in the future.3
Study LimitationsThis was a retrospective study without cause-specific mortal-ity. Diabetes mellitus was defined as known diabetic status onadmission. It is well known that a number of STEMI patientshave undetected diabetes mellitus, and they were not ex-cluded in our study.15,26 Indeed, when the HbA1c cutoff valueof �6.5% as suggested by the American Diabetes Associa-tion was used, �5% of our population could readily bediagnosed with diabetes mellitus on admission (undiagnoseddiabetes mellitus), and they were not excluded from the mainanalysis.27 However, because there was a stepwise increase inmortality with increasing HbA1c levels starting from thelowest to the highest quartile, we believe that our findings arenot solely attributable to the number of patients with unde-tected diabetes mellitus. It is conceivable that some subjectswithin the higher HbA1c quartiles may have progressed toovert diabetes mellitus within the follow-up period, whichwould also adversely affect long-term prognosis. Because wehad no data on the occurrence of this diagnosis or the start ofglucose-lowering medication during the follow-up period, themagnitude of this phenomenon was unknown in our study.Another limitation might be that although admission glucosewill be responsive to the acute stress associated with anSTEMI, many other factors may contribute to the variabilityof nonfasting glucose levels.
ConclusionsBoth elevated admission HbA1c and glucose were associatedwith an adverse prognosis in nondiabetic patients with ST-elevation myocardial infarction treated with primary PCI. TheHbA1c and glucose levels reflect different patient popula-tions, and their association with outcome is probably due todifferent mechanisms. High admission glucose is associatedwith a more hemodynamically unstable patient group with alarger infarct size and high early mortality. Elevated HbA1c isassociated with more adverse baseline characteristics and amore gradual higher mortality over time. Close follow-up ofthese patients seems warranted. More research is needed tobetter describe and understand these findings, and, moreimportant, to assess and develop feasible treatment options.
AcknowledgmentsVera Derks is credited for the submission process.
DisclosuresNone.
References1. Timmer JR, Ottervanger JP, de Boer MJ, Boersma E, Grines CL,
Westerhout CM, Simes RJ, Granger CB, Zijlstra F. Primary percutaneous
coronary intervention compared with fibrinolysis for myocardialinfarction in diabetes mellitus: results from the Primary Coronary Angio-plasty vs Thrombolysis-2 trial. Arch Intern Med. 2007;167:1353–1359.
2. Stolar M. Glycemic control and complications in type 2 diabetes mellitus.Am J Med. 2010;123:S3–S11.
3. Selvin E, Steffes MW, Zhu H, Matsushita K, Wagenknecht L, Pankow J,Coresh J, Brancati FL. Glycated hemoglobin, diabetes, and cardiovascu-lar risk in nondiabetic adults. N Engl J Med. 2010;362:800–811.
4. Goyal A, Mehta SR, Diaz R, Gerstein HC, Afzal R, Xavier D, Liu L, PaisP, Yusuf S. Differential clinical outcomes associated with hypoglycemiaand hyperglycemia in acute myocardial infarction. Circulation. 2009;120:2429–2437.
5. Capes SE, Hunt D, Malmberg K, Gerstein HC. Stress hyperglycaemia andincreased risk of death after myocardial infarction in patients with andwithout diabetes: a systematic overview. Lancet. 2000;355:773–778.
6. Kosiborod M, Rathore SS, Inzucchi SE, Masoudi FA, Wang Y, HavranekEP, Krumholz HM. Admission glucose and mortality in elderly patientshospitalized with acute myocardial infarction: implications for patientswith and without recognized diabetes. Circulation. 2005;111:3078–3086.
7. Kosiborod M. Blood glucose and its prognostic implications in patientshospitalised with acute myocardial infarction. Diab Vasc Dis Res. 2008;5:269–275.
8. Hadjadj S, Coisne D, Mauco G, Ragot S, Duengler F, Sosner P, Tor-remocha F, Herpin D, Marechaud R. Prognostic value of admissionplasma glucose and HbA in acute myocardial infarction. Diabet Med.2004;21:305–310.
9. Cakmak M, Cakmak N, Cetemen S, Tanriverdi H, Enc Y, Teskin O, KilicID. The value of admission glycosylated hemoglobin level in patientswith acute myocardial infarction. Can J Cardiol. 2008;24:375–378.
10. Tenerz A, Nilsson G, Forberg R, Ohrvik J, Malmberg K, Berne C,Leppert J. Basal glucometabolic status has an impact on long-termprognosis following an acute myocardial infarction in non-diabeticpatients. J Intern Med. 2003;254:494–503.
11. Thygesen K, Alpert JS, White HD, Jaffe AS, Apple FS, Galvani M, KatusHA, Newby LK, Ravkilde J, Chaitman B, Clemmensen PM, Dellborg M,Hod H, Porela P, Underwood R, Bax JJ, Beller GA, Bonow R, Van derWall EE, Bassand JP, Wijns W, Ferguson TB, Steg PG, Uretsky BF,Williams DO, Armstrong PW, Antman EM, Fox KA, Hamm CW, OhmanEM, Simoons ML, Poole-Wilson PA, Gurfinkel EP, Lopez-Sendon JL,Pais P, Mendis S, Zhu JR, Wallentin LC, Fernandez-Aviles F, Fox KM,Parkhomenko AN, Priori SG, Tendera M, Voipio-Pulkki LM, VahanianA, Camm AJ, De Caterina R, Dean V, Dickstein K, Filippatos G, Funck-Brentano C, Hellemans I, Kristensen SD, McGregor K, Sechtem U, SilberS, Tendera M, Widimsky P, Zamorano JL, Morais J, Brener S, HarringtonR, Morrow D, Lim M, Martinez-Rios MA, Steinhubl S, Levine GN,Gibler WB, Goff D, Tubaro M, Dudek D, Al-Attar N. Universal defi-nition of myocardial infarction. Circulation. 2007;116:2634–2653.
12. Van de Werf F, Bax J, Betriu A, Blomstrom-Lundqvist C, Crea F, FalkV, Filippatos G, Fox K, Huber K, Kastrati A, Rosengren A, Steg PG,Tubaro M, Verheugt F, Weidinger F, Weis M. Management of acutemyocardial infarction in patients presenting with persistent ST-segmentelevation: the Task Force on the Management of ST-Segment ElevationAcute Myocardial Infarction of the European Society of Cardiology. EurHeart J. 2008;29:2909–2945.
13. The Thrombolysis in Myocardial Infarction (TIMI) trial: phase I findings:TIMI Study Group. N Engl J Med. 1985;312:932–936.
14. Henriques JP, Zijlstra F, van ’t Hof AWJ, de Boer MJ, Dambrink JHE,Gosselink ATM, Hoorntje JCA, Suryapranata H. Angiographicassessment of reperfusion in acute myocardial infarction by myocardialblush grade. Circulation. 2003;107:2115–2119.
15. Norhammar A, Tenerz A, Nilsson G, Hamsten A, Efendic S, Ryden L,Malmberg K. Glucose metabolism in patients with acute myocardialinfarction and no previous diagnosis of diabetes mellitus: a prospectivestudy. Lancet. 2002;359:2140–2144.
16. Timmer JR, van der Horst I, Ottervanger JP, Henriques JP, Hoorntje JCA,de Boer MJ, Suryapranata H, Zijlstra F. Prognostic value of admissionglucose in non-diabetic patients with myocardial infarction. Am Heart J.2004;148:399–404.
17. DECODE Study Group. Is the current definition for diabetes relevant tomortality risk from all causes and cardiovascular and noncardiovasculardiseases? Diabetes Care. 2003;26:688–696.
18. Coutinho M, Gerstein HC, Wang Y, Yusuf S. The relationship betweenglucose and incident cardiovascular events: a metaregression analysis ofpublished data from 20 studies of 95,783 individuals followed for 124years. Diabetes Care. 1999;22:233–240.
710 Circulation August 9, 2011
by IMED Korea on November 13, 2011http://circ.ahajournals.org/Downloaded from
20. Khaw KT, Wareham N, Bingham S, Luben R, Welch A, Day N. Asso-ciation of hemoglobin A1c with cardiovascular disease and mortality inadults: the European Prospective Investigation Into Cancer in Norfolk.Ann Intern Med. 2004;141:413–420.
21. Ryden L, Standl E, Bartnik M, Van den Berghe G, Betteridge J, de BoerMJ, Cosentino F, Jonsson B, Laakso M, Malmberg K, Priori S, OstergrenJ, Tuomilehto J, Thrainsdottir I, Vanhorebeek I, Stramba-Badiale M,Lindgren P, Qiao Q, Priori SG, Blanc JJ, Budaj A, Camm J, Dean V,Deckers J, Dickstein K, Lekakis J, McGregor K, Metra M, Morais J,Osterspey A, Tamargo J, Zamorano JL, Deckers JW, Bertrand M, Char-bonnel B, Erdmann E, Ferrannini E, Flyvbjerg A, Gohlke H, Juanatey JR,Graham I, Monteiro PF, Parhofer K, Pyorala K, Raz I, Schernthaner G,Volpe M, Wood D. Guidelines on diabetes, pre-diabetes, and cardiovas-cular diseases: executive summary: the Task Force on Diabetes andCardiovascular Diseases of the European Society of Cardiology (ESC)and of the European Association for the Study of Diabetes (EASD). EurHeart J. 2007;28:88–136.
22. Orchard TJ, Temprosa M, Goldberg R, Haffner S, Ratner R, MarcovinaS, Fowler S. The effect of metformin and intensive lifestyle intervention
on the metabolic syndrome: the Diabetes Prevention Program randomizedtrial. Ann Intern Med. 2005;142:611–619.
23. Timmer JR, Svilaas T, Ottervanger JP, Henriques JP, Dambrink JHE, vanden Broek SA, van der Horst I, Zijlstra F. Glucose-insulin-potassiuminfusion in patients with acute myocardial infarction without signs ofheart failure: the Glucose-Insulin-Potassium Study (GIPS)-II. J Am CollCardiol. 2006;47:1730–1731.
24. Malmberg K, Ryden L, Wedel H, Birkeland K, Bootsma A, Dickstein K,Efendic S, Fisher M, Hamsten A, Herlitz J, Hildebrandt P, MacLeod K,Laakso M, Torp-Pedersen C, Waldenstrom A. Intense metabolic controlby means of insulin in patients with diabetes mellitus and acute myo-cardial infarction (DIGAMI 2): effects on mortality and morbidity. EurHeart J. 2005;26:650–661.
25. Diaz R, Goyal A, Mehta SR, Afzal R, Xavier D, Pais P, Chrolavicius S,Zhu J, Kazmi K, Liu L, Budaj A, Zubaid M, Avezum A, Ruda M, YusufS. Glucose-insulin-potassium therapy in patients with ST-segment ele-vation myocardial infarction. JAMA. 2007;298:2399–2405.
26. Knudsen EC, Seljeflot I, Abdelnoor M, Eritsland J, Mangschau A,Arnesen H, Andersen GO. Abnormal glucose regulation in patients withacute ST-elevation myocardial infarction: a cohort study on 224 patients.Cardiovasc Diabetol. 2009;8:6.
27. Diagnosis and classification of diabetes mellitus. Diabetes Care. 2010;33(suppl 1):S62–S69.
CLINICAL PERSPECTIVEMeasurement of admission glucose and hemoglobin A1c (HbA1c) in acute myocardial infarction may identify patients withdisturbed glucose metabolism and an increased risk for adverse outcome. Although HbA1c and glucose are related, theycan differentiate between mechanisms of adverse outcome. Admission glucose is related to increased hemodynamic stress,whereas HbA1c identifies patients with higher long-term cardiovascular risk, possibly by abnormal long-term glucoselevels. Early identification of these patient groups enables the initiation of specific intervention strategies and may help usdevelop strategies to improve prognosis in these high-risk patient groups. This is of particular importance because thereis a global increase in the number of patients suffering from cardiovascular disease with underlying insulin resistance,prediabetes, and overt diabetes mellitus. Both glucose and HbA1c should be measured in patients admitted withST-segment–elevation myocardial infarction.
Go to http://cme.ahajournals.org to take the CME quiz for this article.
Timmer et al Glycometabolic Derangements in Myocardial Infarction 711
by IMED Korea on November 13, 2011http://circ.ahajournals.org/Downloaded from