Fasting glucose levels elevated beyond the normal range have been associated with increased cardiovascular risk. However, it is unknown whether this association exists for variations of fasting glucose within the normal range. The present study was conducted using the computerized database of the Sharon-Shomron District of Clalit Health Services. Included in the present study were subjects with fasting glucose levels within the normal range (<100 mg/dl). We excluded patients with a history of cardiovascular disease or diabetes. The primary outcome was the incidence of coronary revascularization with either percutaneous coronary intervention or coronary artery bypass grafting. The 28,263 participants (age 53.7 ± 12.2 years) were divided into quartiles according to the fasting glucose level (75.4 ± 4.5, 83.6 ± 1.7, 88.9 ± 1.4, and 95.1 ± 2.2 mg/dl). During a mean follow-up of 5.9 ± 0.7 years, 424 subjects required coronary revascularization. A progressive increase was seen in the risk of coronary revascularization as the fasting glucose levels increased within the normal range (hazard ratio 1.73, 95% confidence interval 1.3 to 2.3, p >0.001, between the fourth and first quartiles). However, this association lost its statistical significance after adjustments for the conventional coronary risk factors (hazard ratio 1.17, 95% confidence interval 0.85 to 1.62, p = 0.328). In conclusion, elevated fasting glucose levels within the normal range were associated with an increased cardiovascular risk. This association was caused by the greater prevalence of the other conventional risk factors and not by the glucose level itself.
The question of whether elevated fasting plasma glucose levels within the normal range can be used as a marker for increased cardiovascular risk in the nondiabetic population and whether this association acts as an independent risk factor has not been yet answered. The evaluation of whether elevated fasting glucose levels within the normal range are a cardiovascular risk factor is important in public health terms because the increase in the prevalence of obesity has been associated with an upward shift in the population distribution of fasting glucose levels. In the present study, we evaluated the association between fasting plasma glucose levels within the normal range and cardiovascular risk in a large cohort of ambulatory nondiabetic subjects.
Methods
Our cohort included members of the Clalit Health Services, the largest health maintenance organization in Israel. Clalit Health Services insures >50% of the population of the Sharon-Shomron District, located in central Israel. The district’s population is mostly urban and includes both Jews and Arabs. All medical information from the primary care clinics is recorded in the Clalit Health Services’ computerized database and can be accessed at the level of the individual patient. Each family physician is responsible for updating the computerized medical records routinely during all visits to the primary care clinic and after any admission to one of the regional hospitals. The database includes a list of all diagnoses, demographic data, laboratory values, medications, and medical procedures. This database was the source of data for our study.
Men and women ≥30 years old who were registered in Clalit Health Services database and were living in the Sharon-Shomron district in 2002 and for whom fasting blood glucose levels were available during that year were included in the present cohort. Because the target of the study was to evaluate the association of fasting plasma glucose with cardiovascular outcomes in a healthy population, we excluded patients with a diagnosis of diabetes mellitus, dyslipidemia, hypertension, or any cardiovascular disease (including coronary artery disease, ischemic or hemorrhagic stroke, transient ischemic attack, peripheral vascular disease, and heart failure). Those patients receiving long-term cardiovascular (including antiplatelet, anticoagulant, antihypertensive, lipid-lowering, or antiarrhythmic drugs) or antidiabetic medications were also excluded.
The incidence of coronary revascularization with either percutaneous coronary intervention (PCI) or coronary artery bypass grafting (CABG) was the primary end point of our study. The revascularization procedures are performed at the regional hospitals and are automatically documented in the Clalit Health Services database independently of the family physician’s follow-up. This has made them the most reliable cardiovascular end points in our database; therefore, they were chosen as the primary end points.
Hypertension was considered when such a diagnosis appeared on the patient’s diagnosis list and/or when long-term treatment with antihypertensive medications was administered. Specific blood pressure measurements were available only for a minority of the population and therefore were not included in the present study. Patients were considered to have diabetes mellitus if this diagnosis appeared in their medical records or when regular treatment with oral hypoglycemic agents or insulin was administered. Dyslipidemia was defined as low-density lipoprotein (LDL) cholesterol levels that required medical treatment according to the Adult Treatment Panel III guidelines or if long-term treatment with lipid-lowering drugs was given. The body mass index (BMI) was calculated as the weight in kilograms divided by the square of the height in meters. The height and weight were measured at the primary care clinics.
The biochemical analyses of blood samples, including the measurement of serum glucose and lipids, were performed on fresh samples at a core laboratory facility. The laboratory is authorized to perform tests according to the International Organisation for Standardization (ISO)-9002. Periodic assessment of quality control was regularly performed. All measured biochemical markers were identified using a BM/Hitachi917 automated analyzer (Boehringer, Mannheim, Germany).
The local institutional ethics committee approved the study in keeping with the principles of the Declaration of Helsinki. In accordance with Ministry of Health regulations, the institutional ethics committee did not require written informed consent, because the data had been collected anonymously from the computerized medical files, with no participation by the patients.
The 28,263 study subjects were stratified into quartiles according to their fasting glucose level determined at entrance into the study. The baseline clinical characteristics of these 4 patient groups were compared using analysis of variance. Cox proportional hazards regression analysis with the years of follow-up as the time scale was used to estimate the hazard ratios and 95% confidence intervals for coronary revascularization (PCI or CABG) according to the fasting plasma glucose quartile. This analysis was adjusted for age, gender, LDL and high-density lipoprotein cholesterol level, BMI, and smoking. We also conducted a separate subanalysis according to gender. The statistical analysis was performed using the Statistical Package for Social Sciences, for Windows, version 17.0 (SPSS, Chicago, Illinois).
Results
The database included 154,914 patients >30 years old for whom fasting glucose levels were available during 2002. Of these patients, 126,651 were excluded because of abnormal fasting glucose levels (≥100 mg/dl), because of a diagnosis of diabetes mellitus, dyslipidemia, or any cardiovascular disease, or because of long-term treatment with cardiovascular or antidiabetic medications, according to our exclusion criteria. A total of 28,263 subjects with a mean age of 53.7 ± 12.2 years were available for analysis. The population consisted of 10,937 men (38.7%) and 17,326 women (61.3%), and the mean fasting glucose level was 85.9 ± 7.7 mg/dl.
When we divided the cohort into quartiles according to the fasting glucose levels, we noted that those with the greater glucose levels within the normal range were older, more frequently men, and included a greater proportion of smokers. In addition, the greater the glucose level, the greater the BMI and total, LDL cholesterol, and triglycerides levels and the lower the high-density lipoprotein cholesterol levels ( Table 1 ).
| Variable | Quartile | p Value | |||
|---|---|---|---|---|---|
| 1 (n = 6,748) | 2 (n = 7,609) | 3 (n = 6,481) | 4 (n = 7,425) | ||
| Mean follow-up (years) | 5.86 ± 0.74 | 5.87 ± 0.7 | 5.87 ± 0.7 | 5.84 ± 0.77 | 0.5 |
| Fasting blood glucose level (mg/dl) | |||||
| Mean | 75.4 ± 4.5 | 83.6 ± 1.7 | 88.9 ± 1.4 | 95.1 ± 2.2 | |
| Median | 77 | 84 | 89 | 95 | |
| Range | ≤80 | 81–86 | 87–91 | 92–96 | |
| Age (years) | 51.8 ± 12.1 | 52.9 ± 12.1 | 54.2 ± 11.9 | 55.9 ± 12.3 | <0.001 |
| Men | 2,277 (33.7%) | 2,769 (36.4%) | 2,594 (40%) | 3,297 (44%) | <0.001 |
| Total cholesterol (mg/dl) | 202 ± 39 | 206 ± 39 | 206 ± 38 | 209 ± 39 | <0.001 |
| Low-density lipoprotein cholesterol (mg/dl) | 125 ± 33 | 128 ± 33 | 129 ± 32 | 131 ± 33 | <0.001 |
| High-density lipoprotein cholesterol (mg/dl) | 53 ± 14 | 52 ± 14 | 51 ± 13 | 50 ± 13 | <0.001 |
| Triglycerides (mg/dl) | 121 ± 72 | 126 ± 73 | 131 ± 82 | 140 ± 87 | <0.001 |
| Body mass index (kg/m 2 ) | 26 ± 5 | 27 ± 5 | 27 ± 5 | 28 ± 5 | <0.001 |
| Current smokers (%) | 18.7% | 18.7% | 19.5% | 21% | 0.004 |
During the study period, 424 subjects (1.5%) underwent 440 revascularization procedures with either PCI (n = 344, 78%) or CABG (n = 96, 22%); 16 underwent both procedures. We observed a significant and progressive increase in the risk of coronary revascularization with an increasing fasting glucose level within the normal range ( Table 2 ). The risk of coronary revascularization was significantly greater for the fourth than for the first quartile (hazard ratio 1.73, 95% confidence interval 1.3 to 2.3, p <0.001; Table 2 ). This association remained significant after adjustment for age (hazard ratio 1.48, 95% confidence interval 1.1 to 1.98, p = 0.011; Table 2 ). However, using a multivariate model, adjusted for age, gender, cholesterol level, BMI, and smoking status, the association between the fasting glucose level within the normal range and the risk of coronary revascularization lost its statistical significance (hazard ratio 1.17, 95% confidence interval of 0.85 to 1.62, p = 0.328). The conventional risk factors, including age (p <0.001), male gender (p <0.001), cigarette smoking (p <0.001), LDL cholesterol level (p <0.001), high-density lipoprotein cholesterol level (p = 0.001), and BMI (p = 0.002), remained independent predictors of coronary disease even in the multivariate model.
