Association of Hemoglobin A1c with Cardiovascular Disease and Mortality in Adults: The European Prospective Investigation into Cancer in Norfolk

Context Several studies suggest that blood glucose levels are associated with cardiovascular disease, even at blood glucose values that do not meet diagnostic criteria for diabetes. Contribution Among adult residents of Norfolk, United Kingdom, there was a continuous relationship between hemoglobin A1c levels and cardiovascular disease and total mortality. This relationship was apparent even among persons without diabetes. Implications These observations justify the need for studies that address whether improvements in glycemic control might improve health outcomes in persons who do not have diabetes. The Editors Diabetes mellitus is of major and increasing global public health importance (1). Persons with diabetes are at increased risk for premature disability and death associated with vascular, renal, retinal, and neuropathic complications. Raised fasting and postchallenge blood glucose levels in an oral glucose tolerance test are used to diagnose diabetes. The diagnostic threshold is based on the shape of the risk curve between glucose levels and specific microvascular complications of diabetes (2-6). Diabetes also increases the risk for macrovascular diseases, such as coronary heart disease and stroke (7). In contrast to microvascular disease, increasing evidence suggests that the relationship between blood glucose level and macrovascular disease is continuous and does not have an obvious threshold (2, 8, 9). Hemoglobin A1c concentration is an indicator of average blood glucose concentrations over the preceding 3 months; it is useful for characterizing dysglycemia in population studies because it is simpler to perform than the oral glucose tolerance test (10). In a 3-year follow-up of men in a prospective study, we previously reported that hemoglobin A1c concentrations were related to cardiovascular disease and all-cause mortality (11). However, we had insufficient power to examine risk relationships at concentrations close to the diagnostic threshold of 7% or to examine the relationship in women. We report the relation between hemoglobin A1c concentrations and fatal and nonfatal coronary heart disease, cardiovascular disease events, and all-cause mortality in men and women after an average of 6 years of follow-up. Methods The European Prospective Investigation into Cancer in Norfolk (EPICNorfolk) is a prospective population study of 25 623 men and women who were between 40 and 79 years of age and who resided in Norfolk, United Kingdom. Participants were recruited from general practice registers. Information on the recruitment process is available elsewhere (12). Between 1993 and 1997, participants completed a health and lifestyle questionnaire. Participants were asked whether a doctor had ever told them that they have any of the conditions contained in a list that included diabetes, heart attack, and stroke. People with known diabetes were defined as those who responded yes to the diabetes option of this question. Smoking history was derived from responses (yes or no) to the questions: Have you ever smoked as much as 1 cigarette a day for as long as a year? and Do you smoke cigarettes now? At a clinic, trained nurses performed a health examination for each participant. Body mass index was estimated as weight (kg)/height (m2), and waist-to-hip ratio was determined by measurements of the circumference of the waist and hips. Blood pressure was measured by using an Accutorr (Datascope, Mahwah, New Jersey) noninvasive blood pressure monitor after the participant had been seated for 5 minutes. The mean of 2 readings was used for analysis. Nonfasting blood samples were taken; samples for assay were stored in a refrigerator at 4 C until transport within 1 week of sampling to the Department of Clinical Biochemistry, University of Cambridge. Starting in 1995, hemoglobin A1c was measured on fresh EDTA blood samples by using high-performance liquid chromatography (BioRad Diamat Automated Glycosylated Haemoglobin Analyser, Hemel Hempstead, United Kingdom). We report results for follow-up to January 2003, an average of about 6 years. All participants were flagged for death certification at the Office of National Statistics; vital status was obtained for the entire cohort. Trained nosologists coded death certificates according to the International Classification of Diseases, Ninth or Tenth Revisions (ICD-9 or ICD-10). Cardiovascular death (stroke, coronary heart disease, and other vascular causes) was defined as those whose underlying cause of death was coded as ICD-9 400448 or ICD-10 I10I79. Death from coronary heart disease was defined as those whose cause of death was coded as ICD-9 410414 or ICD-10 I22I25. Participants admitted to a hospital were identified by their National Health Service number. Hospitals were linked to the East Norfolk Health Authority database, which identifies all hospital contacts throughout England and Wales for Norfolk residents. We used the same ICD diagnostic codes described in the preceding paragraphs to ascertain hospital episodes of cardiovascular disease and coronary heart disease in our cohort. Participants were identified as having a coronary heart disease event during follow-up if they had a hospital admission or died with coronary heart disease as the cause of death. Of the coronary heart disease events identified, 21% (112 of 529) were fatal; of the cardiovascular disease events, 23% (197 of 806) were fatal. In men, 24% (76 of 321) of deaths were attributed to heart disease and 29% (117 of 321) were attributed to cardiovascular disease. In women, 18% (36 of 200) of deaths were attributed to heart disease and 35% (70 of 200) were attributed to cardiovascular causes. The Norwich Ethics Committee approved the study, and participants gave signed informed consent. Statistical Analysis These analyses, undertaken by using SPSS software, version 10.0 (SPSS, Inc., Chicago, Illinois), included 10 232 men and women age 45 to 79 years who completed the health and lifestyle questionnaire and had available hemoglobin A1c measurements. We divided the cohort into 7 categories on the basis of baseline data: known diabetes, high likelihood of previously undiagnosed diabetes (no personal history of diabetes but a hemoglobin A1c concentration 7%), and hemoglobin A1c concentrations in 0.5percentage point intervals (<5%, 5% to 5.4%, 5.5% to 5.9%, 6.0% to 6.4%, and 6.5% to 6.9%). We examined risk factor distributions and then coronary heart disease, cardiovascular disease, and all-cause mortality rates by hemoglobin A1c and diabetes category. Age-adjusted odds ratios were calculated by using logistic regression models. We used a Cox proportional hazards model to determine the independent contribution of hemoglobin A1c to total mortality and cardiovascular and coronary heart disease after adjustment for age, body mass index, waist-to-hip ratio, systolic blood pressure, blood cholesterol concentrations, cigarette smoking, and history of heart attack or stroke. Participants with missing baseline data for 1 or more risk factors (130 men and 186 women) were excluded from the multivariate analyses. Role of the Funding Sources The funding sources had no role in the design, conduct, and reporting of the study or in the decision to submit the manuscript for publication. Results Table 1 presents characteristics of the participants according to hemoglobin A1c concentration and self-reported diabetes. Those with known diabetes had higher mean (SD) hemoglobin A1c concentrations (8.0% 1.9%) than the rest of the study sample (5.3% 0.7%). They were older and had a higher body mass index, waist-to-hip ratio, and systolic blood pressure; they were also more likely to report having had a previous heart attack or stroke. Participants with probable but previously undiagnosed diabetes (hemoglobin A1c 7%) shared these characteristics. Mean risk factor levels rose with increasing concentration of hemoglobin A1c less than 7%. Table 1. Distribution of Variables by Hemoglobin A1c Concentration and Known Diabetes in 4662 Men and 5570 Women Age 45 to 79 Years (European Prospective Investigation into Cancer in Norfolk, 1995 to 1997) Table 2 shows adjusted odds ratios for hemoglobin A1c concentrations, diabetes status, and outcomes. Persons with known or undiagnosed diabetes had a greater risk for all-cause mortality and cardiovascular or coronary heart disease than those without diabetes. Risk for coronary heart or cardiovascular disease and total mortality increased throughout the whole range of hemoglobin A1c concentrations; those with hemoglobin A1c concentrations less than 5% had the lowest rates. For men, a gradient of increasing rates through the distribution was apparent for all end points. For women, odds ratios for cardiovascular or coronary heart disease did not increase significantly until the hemoglobin A1c concentration reached 6%; odds ratios were very high in women with concentrations greater than 7%. Table 2. Rates and Age-Adjusted Relative Risks for Total Coronary Heart Disease Events, Cardiovascular Disease Events, and All-cause Mortality by Category of Hemoglobin A1c Concentration and Known Diabetes in 4462 Men and 5570 Women Age 45 to 79 Years (European Prospective Investigation into Cancer in Norfolk, 1995 to 2003) Table 3 shows outcomes after adjustment for age alone and then after adjustment for age and other risk factors. In men, known diabetes predicted coronary heart and cardiovascular disease events and total mortality with approximate 2-fold relative risks. These relative risks were only slightly attenuated after adjustment for known risk factors. In women, known diabetes status predicted an approximate 5-fold increase in risk for coronary heart and 3-fold increase in risk for cardiovascular disease events; these increases were attenuated after adjustment for known risk factors to 3-fold and 2-fold risk, respectively. In men and women, hemoglobin A1c concentrations predicted an increased risk for coronar