|Home | About | Journals | Submit | Contact Us | Français|
To examine the association between hyperglycemia and subclinical myocardial injury in persons without clinically evident coronary heart disease (CHD).
Hyperglycemia is associated with an increased risk of cardiac events but limited information is available on its relationship to subclinical myocardial damage. Elevated cardiac troponin T even below traditional detection levels can be detected by a novel high-sensitivity assay.
We examined the association between baseline glycated hemoglobin (HbA1c) and high-sensitivity cardiac troponin T (hs-cTnT) in 9,661 participants free of CHD and heart failure in the Atherosclerosis Risk in Communities study. Multivariable logistic regression models characterized the association between clinical categories of HbA1c (<5.7, 5.7–6.4 and ≥6.5%) and our primary outcome of elevated hs-cTnT (≥ 14 ng/L).
Higher baseline values of HbA1c were associated in a graded fashion with elevated hs-cTnT (P-for-trend <0.001). After adjusting for traditional risk factors, compared to persons with HbA1c <5.7%, the odds ratio of elevated hs-cTnT for persons with HbA1c 5.7–6.4% and ≥6.5% was 1.26 (95% CI 1.01, 1.56) and 1.97 (1.44, 2.70), respectively.
Higher HbA1c is associated with elevated hs-cTnT among persons without clinically evident CHD, suggesting that hyperglycemia contributes to myocardial injury beyond its effects on development of clinical atherosclerotic coronary disease.
Cardiovascular disease (CVD) is the leading cause of morbidity and mortality in persons with diabetes mellitus(1). Although the excess cardiovascular risk observed in dysglycemic states has been partially explained by coexisting cardiovascular risk factors (e.g. obesity and hypertension), there is strong evidence supporting hyperglycemia, even below the threshold of diabetes, as an independent risk factor for coronary heart disease (CHD) morbidity and mortality(2). Glycated hemoglobin (HbA1c), a marker of chronic hyperglycemia, is associated with incident CHD(3,4) and all-cause mortality(4–6) and has recently been recommended as a diagnostic test for diabetes mellitus(7).
Much less is known about the association of chronic hyperglycemia with subclinical CHD. A small number of studies have used imaging techniques to examine the cross-sectional relationship of HbA1c with subclinical atherosclerosis(8,9). There is limited information on the relationship between hyperglycemia and subclinical myocardial injury at earlier stages, before it can be detected by imaging modalities.
Newly identified biomarkers allow for the noninvasive identification of subclinical myocardial damage(10). Cardiac Troponin-T (cTnT) has been independently associated with CHD (11,12). A novel high-sensitivity cardiac troponin-T (hs-cTnT) assay allows for the detection of hs-cTnT levels below clinical thresholds or detection levels of earlier assays. Using this novel assay, troponin levels below the conventional limit of detection have recently been shown to predict cardiovascular events and death in persons with CHF(13), CAD(14) and in the general population(15–17). Some have suggested that minimally elevated troponin levels may represent subclinical cardiac injury(18,19) and hs-cTnT may now allow for the identification of previously undetected early stages of myocardial injury.
The objective of this study was to test the hypothesis that in persons without clinically evident CHD or heart failure, chronic hyperglycemia--assessed by HbA1c--would be independently associated with subclinical myocardial injury measured by elevated hs-cTnT.
More detailed methods are described in the online Supplement. Briefly, we explored the association between categories of HbA1c and hs-cTnT in participants without CHD in the Atherosclerosis Risk in Communities Study. Hs-cTnT was measured using a novel high-sensitivity assay, Elecsys Troponin T (Roche Diagnostics, Indianapolis, IN) with a lower limit of detection of 3.0 ng/L. Elevated hs-cTnT was defined as levels above the previously reported 99th percentile value (14 ng/L) in a healthy subpopulation aged 20–70 years.
HbA1c was modeled as a continuous variable and by clinical categories: <5.7, 5.7–6.4 and ≥ 6.5%(7) and baseline characteristics were compared across these categories. Linear splines (with knots at clinical cut-points) were used to characterize the shape of the association between HbA1c and hs-cTnT and multivariable logistic regression to assess the association between categories of HbA1c and elevated hs-cTnT (≥14.0 ng/L). For continuous models, undetectable levels of hs-cTnT were assigned a value of 1.5 ng/L (i.e., half the lower limit of detection). In participants who fasted ≥8 hours at the time of visit 2 (n=9,550), we evaluated the association between categories of fasting glucose: <100, 100–125, ≥126 mg/dL.
We implemented four models for the adjustment of covariates. Model 1 unadjusted; Model 2 adjusted for age, gender and race; Model 3 was adjusted for field center, body mass index, education, systolic and diastolic blood pressures, smoking, LDL- and HDL-cholesterol, alcohol, hypertension medication, left ventricular hypertrophy, history of diabetes and eGFR; Model 4a was adjusted for fasting glucose and Model 4b for HbA1c.
Sensitivity analyses excluded persons with atrial fibrillation, eGFR <60 ml/min/1.73 m2 or stroke, before or at visit 4. We also excluded persons with incident CHD up to 6 months after hs-cTnT samples were obtained. We tested for interaction with age, race, and sex. Analyses were conducted using STATA 11.1 (Stata Corp, College Station, TX) and a P-value of <0.05 was statistically significant.
Mean age was 56.6 years and 58.9% were female (Table 1). Participants with higher HbA1c were more likely to be African American, hypertensive and have higher BMI, LDL and lower HDL-cholesterol. A total of 6,400 individuals (66.3%) had detectable levels of hs-cTnT. In persons with and without a history of diagnosed diabetes, the corresponding percentages were 81.9% and 65.2%.
Baseline HbA1c was positively associated with hs-cTnT in a linear fashion (tests for differences in the slope at clinical cut-points of HbA1c were significant) (Figure 1). Compared to persons with HbA1c <5.7%, hs-cTnT values were ~25% higher in persons with HbA1c 5.7–6.4% and ~70% higher among participants with HbA1c ≥6.5%. In fully adjusted models, every 1-percentage point higher HbA1c value was associated with a 0.7 ng/L higher value of hs-cTnT (95%CI=0.5, 1.0; P< 0.001)
Overall, 7.3% of participants had elevated hs-cTnT (≥14 ng/L). In persons with and without diabetes, the corresponding percentages were 21.8% and 6.3%. In this population of persons without clinically evident CHD or heart failure, the 99th percentile of hs-cTnT was 30 ng/L. The corresponding 99th percentiles were 27 ng/L and 54 ng/L for persons without and with a history of diabetes, respectively. Increasing HbA1c categories were associated in a graded fashion with elevated hs-cTnT (Figure 2). In analysis adjusted for age, gender and race, the percentage of persons with elevated hs-cTnT among persons with a HbA1c <5.7, 5.7–6.4 and ≥6.5% was 3.8%, 5.7% and 14.0%, respectively.
Compared to persons with HbA1c <5.7%, higher HbA1c values were strongly associated with elevated hs-cTnT (Table 2). In age, race and gender adjusted analyses, increasing HbA1c categories was significantly associated with elevated hs-cTnT for HbA1c of 5.7–6.4% (OR 1.51; 95% CI 1.23, 1.85) and for HbA1c ≥6.5% (OR 4.09; 95% CI 3.25, 5.15). After adjusting for other cardiovascular risk factors, the association was attenuated but remained significant. This association was unaltered by further adjustment for fasting glucose (Table 2, Model 4a). Analysis performed stratified by previously diagnosed diabetes demonstrated a similar association (supplementary table 1). Excluding participants with eGFR <60 ml/min/1.73 m2, stroke, atrial fibrillation prior to or at visit 4 and persons with incident CHD within 6 months after hs-cTnT measurement did not alter the results (supplementary table 2).
In age, race and gender adjusted analyses, compared to persons with a fasting glucose <100 mg/dL, higher fasting glucose levels were associated with elevated hs-cTnT (OR for glucose ≥126 mg/dL=2.93, 95% CI 2.30–3.75) (Table 3). This association was no longer significant after adjusting for additional cardiovascular risk factors or after further adjustment for HbA1c (Table 3, Model 4b).
There was no interaction between HbA1c and gender or race for elevated hs-cTnT (P>0.110 for both). There was evidence that age may modify the association between HbA1c and elevated hs-cTnT (P-for-interaction=0.048). When stratified by median age (56 years), compared to persons HbA1c <5.7%, HbA1c ≥6.5% was more strongly associated with elevated hs-cTnT among younger (OR=2.13; 95% CI 1.23–3.69) than older individuals (OR= 1.88; 95% CI 1.28–2.76).
In a community-based population of almost 10,000 individuals without clinically evident CHD, chronic hyperglycemia was independently associated with subclinical myocardial injury, as assessed by elevated levels of hs-cTnT in both persons with and without diabetes.
Elevated troponin levels have been previously associated to cardiovascular events(11). Using conventional assays, troponin can be detected in 0.7% of the population(20) and is associated with myocardial infarction and death(11). In contrast, the novel high-sensitivity troponin assay allows for the detection of troponin at levels far below previous detection limits. We observed that hs-cTnT was present in 66.3% of the ARIC population without clinically evident CHD. The 99th percentile value for hs-cTnT in our study (30 ng/L) was considerably higher than that reported by the manufacturer (14 ng/L) in a healthy population aged 20–70 years. Firm reference ranges for this assay have not yet been established.
Others have recently shown in both ARIC(17) and other(15,16) cohorts that troponin levels measured with the highly sensitive assay independently predict cardiovascular events. In ARIC, hs-cTnT levels above the 99th percentile were associated with all-cause mortality (OR 3.69, 95% CI 3.21, 4.88)(17). Also of interest, minimally elevated levels of hs-cTnT (3 to 5 ng/L) were associated with all-cause mortality (OR 1.37, 95% CI 1.14, 1.65).
Numerous studies have demonstrated associations between hyperglycemia and macrovasular and microvascular complications(2–4,6). The use of this high sensitivity assay allowed for the detection of minimally elevated troponin levels thought to represent subclinical cardiac injury(18–20). Although hyperglycemia is traditionally thought to be associated to cardiovascular events via atherosclerosis, our results suggest that hyperglycemia may contribute to cardiovascular events by alternate mechanisms.
A potential mechanism by which hyperglycemia induces myocardial injury is hyperglycemia-mediated coronary microvascular dysfunction(21,22). Other mechanisms include oxidative stress(21), advanced glycation end-products(23,24) and myocardial fibrosis(25). Diabetic subjects have increased oxidative stress(26) and administration of antioxidant vitamin C has been shown to improve endothelial dysfunction in diabetics(27). Finally, hyperglycemia may also be associated with myocardial damage through silent atherosclerotic disease and as others have shown, up to 12% of asymptomatic diabetic subjects had coronary artery stenosis greater than 75%(28).
While some have demonstrated associations of hyperglycemia with subclinical CVD using imaging modalities(8,9), to our knowledge we are the first to demonstrate an association between hyperglycemia and early subclinical myocardial injury. Recent studies using this new highly sensitive assay for cTnT suggest that the association with cardiac outcomes is mediated by mechanisms independent of atherosclerosis(14–17). While left ventricular mass was independently associated with detectable levels of hs-cTnT, coronary artery calcium –a marker of coronary atherosclerosis- was not(16). Hs-cTnT had a stronger association with total mortality and heart failure than to CHD(17). Furthermore, the association with non-fatal CHD was even weaker than for fatal CHD, again suggesting that hs-cTnT is related to outcomes by mechanisms other than atherosclerosis. Nevertheless, due to the observational nature of these investigations, the non-atherosclerotic mechanisms remain unclear.
Our results suggest that HbA1c performs better as a marker of subclinical myocardial damage compared to fasting glucose and are consistent with the growing body of literature demonstrating that HbA1c is an important marker of cardiovascular risk(4). Our data support new recommendations for the use of HbA1c for the diagnosis of diabetes and identification of persons at high risk for development of complications(7).
Our study has important limitations: Measurements of HbA1c and hs-cTnT were not available at the same visit. Although in sensitivity analysis we excluded persons with any history of clinical CHD at visit 4 and incident CHD in the 6 months following hs-cTnT measurement, we cannot rule out the possibility that a number of subjects had elevated troponin levels at the baseline visit. While elevated hs-cTnT levels have now been associated with increased CVD(15–17), whether clinical management should be modified on the basis of chronically elevated levels of troponin is unknown. Because we excluded persons that died or developed CHD between visit 2 and visit 4 and HbA1c is associated with both elevated troponin levels and mortality, there is a possibility that this resulted in selection bias. This would tend to underestimate the true association. Although we adjusted for known risk factors for CHD, we cannot exclude the possibility of residual confounding in this observational study. There were 679 people excluded from our analysis (<10% of the study sample) due to missing data. Nonetheless, this represents one of the largest community-based studies of HbA1c and hs-cTnT. Additional major strengths of this study include the large sample of persons with and without diabetes, including a large number of African Americans, rigorous measurement of cardiovascular risk factors, and our ability to exclude clinical CHD cases utilizing comprehensive and adjudicated surveillance data for clinical events.
In conclusion, in this community-based study of persons without clinically evident CHD, HbA1c was associated with hs-cTnT in a graded fashion. Our findings suggest that hyperglycemia contributes to myocardial injury beyond its effects on development of clinical atherosclerotic coronary disease.
Funding: ARIC is supported by NHLBI contracts N01-HC-55015 through N01-HC-55022. J.R. was supported by NHLBI grant 5T32HL007024, E.S. by NIH/NIDDK grant R21 DK080294 and K01 DK076595, J.C. by NIH/NIDDK grant R01-DK-076770.
Role of the Sponsor: The NIH had no role in the collection, analysis, interpretation of the data or preparation, review and approval of the manuscript.
Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.