Frederick L. Brancati

Meta-analysis: Glycosylated hemoglobin and cardiovascular disease in diabetes mellitus

Context The relationship between glycosylated hemoglobin and cardiovascular disease in diabetic persons is less clear than its relationship with microvascular disease. Contribution This meta-analysis of 13 observational studies estimates that, for every 1-percentage point increase in glycosylated hemoglobin, the relative risk for any cardiovascular disease event is 1.18 for patients with type 2 diabetes mellitus and 1.15 for patients with type 1 diabetes mellitus. Cautions Although this analysis suggests that improvements in glycosylated hemoglobin level might translate into reductions in cardiovascular events, confirmation from randomized trials is necessary. The Editors Persons with diabetes mellitus are at an increased risk for cardiovascular disease; they have more than a 2-fold increased risk for cardiovascular death compared with persons without diabetes (1-3). Cardiovascular death accounts for more than 75% of all deaths among persons with diabetes mellitus (3, 4). Because this excess risk is only partially explained by traditional risk factors, such as obesity, dyslipidemia, and hypertension, diabetes is often considered an independent risk factor for cardiovascular disease. A strong body of evidence links chronic hyperglycemia to microvascular complications, such as retinopathy, neuropathy, and nephropathy, in persons with diabetes (5-10). In randomized clinical trials, improving glycemic control substantially reduces the incidence of microvascular disease in persons with diabetes (5, 6, 11). However, few randomized trials have specifically been designed to examine the influence of glycemic control on macrovascular complications, such as coronary heart disease, stroke, and peripheral arterial disease. Results from clinical trials that collected information on cardiovascular outcomes have been equivocal. In interpreting recent clinical trial data in a position statement, the American Diabetes Association stated that the role of hyperglycemia in cardiovascular complications is still unclear (12). Fasting blood glucose levels in diabetic and nondiabetic persons have been linked to an excess risk for cardiovascular disease (13-15); this link suggests an association between glycemic control and cardiovascular risk. A meta-regression analysis that combined data from more than 95 000 persons without diagnosed diabetes found a graded relationship between fasting and postprandial blood glucose levels and subsequent risk for a cardiovascular event (15). An important clinical question is whether improving long-term glycemic control in persons with diabetes reduces the risk for cardiovascular disease events. Glycosylated hemoglobin reflects long-term glycemic control and is a more accurate and stable measure than fasting blood glucose levels (16). It tracks well over time in persons with diabetes and has less measurement error than fasting blood glucose (17-20). Glycosylated hemoglobin is at the center of the clinical management of hyperglycemia in persons with diabetes. However, clinical guidelines for glycosylated hemoglobin levels are based on cut-points relevant for the prevention of microvascular complications (21). The relationship between glycosylated hemoglobin and cardiovascular disease, the most deadly complication of diabetes mellitus, has not been adequately characterized. We performed a systematic review to characterize the risk relation between long-term glycemic control, as measured by glycosylated hemoglobin, and cardiovascular end points (peripheral arterial disease, coronary heart disease, and cerebrovascular disease) in persons with type 1 and type 2 diabetes mellitus. Methods Study Design We systematically reviewed prospective cohort studies of glycosylated hemoglobin and cardiovascular disease in persons with diabetes mellitus. This study was part of a larger project commissioned by the Agency for Healthcare Research and Quality, which was conducted by the Johns Hopkins Evidenced-based Practice Center (22). Study Selection We searched the MEDLINE database for articles published in English from 1966 to July 2003 by using Medical Subject Heading terms and text words related to cardiovascular disease (coronary heart disease, peripheral arterial disease, or cerebrovascular disease), diabetes mellitus, glycemic control, and glycosylated hemoglobin (the Appendix contains the full text of the search string). We reviewed all abstracts obtained from our search for relevance. We manually reviewed bibliographies and review articles for additional citations and obtained the full text of all potentially relevant articles. We also queried experts to identify any additional studies. Our prespecified inclusion criteria were as follows: 1) prospective cohort studies that examined the cardiovascular outcomes of interest (peripheral arterial disease, coronary heart disease, and stroke) and 2) studies that reported a measure of glycosylated hemoglobin and that were conducted in samples that included persons with type 1 or type 2 diabetes. Persons described as having insulin-dependent diabetes mellitus or younger- or juvenile-onset diabetes were classified as having type 1 diabetes. Individuals described as having noninsulin-dependent diabetes mellitus or older-onset diabetes were classified as having type 2 diabetes. We excluded studies if they 1) had no original data, 2) did not address persons with diabetes, 3) involved nonprospective studies (for example, cross-sectional and retrospective casecontrol studies), 4) had less than 1 year of follow-up, 5) assessed the effect of glycemic control on cardiovascular outcomes after admission to a hospital or after surgery, and 6) involved only patients receiving dialysis or transplants. We excluded 1 additional study (23) in which the outcome was self-reported and the authors did not use standard definitions for classifying cardiovascular outcomes. When several, sequentially published studies were performed in the same sample, the publication with the longest follow-up was selected for inclusion in our analysis. For multiple studies of the same sample with equivalent follow-up, the most recent publication was selected. Data Abstraction Two investigators independently reviewed each article that met the selection criteria and abstracted the data by using standardized data abstraction forms. Discrepancies were resolved by consensus. Data abstracted were age, percentage of male and female study participants, sample size, outcome or outcomes, duration of follow-up, method of measuring glycosylated hemoglobin, main results, statistical methods, number of study participants included in the final analysis, and variables included in the adjusted model or models. For each prospective cohort study that met our inclusion criteria, we abstracted adjusted effect estimates (odds ratios, relative risks, or relative hazards) for the association between cardiovascular risk (based on incident events during follow-up) and baseline or updated mean glycosylated hemoglobin values. Standard errors for the estimates were abstracted or derived by using data reported in the manuscript. The cardiovascular disease end points, defined a priori, were fatal and nonfatal coronary heart disease (myocardial infarction, angina, and ischemic heart disease); cerebrovascular disease (fatal and nonfatal stroke); peripheral arterial disease (lower-extremity peripheral arterial disease, amputation, and claudication); and a combined cardiovascular disease outcome that included studies of coronary heart disease and stroke (but not peripheral arterial disease). We conducted separate analyses for each cardiovascular end point and for samples of persons with type 1 and type 2 diabetes. Studies using a combined outcome that included both coronary heart disease and stroke (24-26) were excluded from the pooled effect estimates for stroke alone and coronary heart disease alone but were included in the combined coronary heart disease and stroke subgroup. Statistical Analysis We conducted separate meta-analyses of the prospective cohort studies for study samples of persons with type 1 and type 2 diabetes and for the different cardiovascular outcomes. Most studies reported glycosylated hemoglobin as percentage hemoglobin A1c or its equivalent, although some studies (27-32) measured hemoglobin A1 and 1 study (33) measured total glycosylated hemoglobin. Although the American Diabetes Association advises that all measurements of glycosylated hemoglobin be reported as percentage hemoglobin A1c or its equivalent (16), there are direct linear relationships between glycosylated hemoglobin subfractions (34); therefore, we did not consider the measured subfraction to be an important source of heterogeneity across studies. For 4 studies (27, 30-32) that reported relative risk estimates for participants in the highest tertile of glycosylated hemoglobin compared with participants in the 2 lowest tertiles, we assumed a normal distribution for glycosylated hemoglobin values and used the reported mean and SD to estimate the 33rd and 83rd percentiles of glycosylated hemoglobin (corresponding to the midpoints of the 2 lowest and the highest tertiles, respectively). Then, we divided the log relative risk by the difference of these 2 values to estimate the effect of a 1-unit change in glycosylated hemoglobin (35). For the study (36) that reported a dichotomous relative risk estimate (it compared glycosylated hemoglobin above and below the median), we used the same method to estimate the effect of a 1-unit change in glycosylated hemoglobin but calculated the 25th and 75th percentiles and divided the log relative risk by the difference of these 2 values. One study (29) did not report relative risks or odds ratios but reported the mean and SD of glycosylated hemoglobin in persons with and without cardiovascular disease events. In this case, we estimated the odds ratio and its 95% CI on the basis of a linear discriminant function model. This m

The prevalence and etiology of elevated aminotransferase levels in the United States.

OBJECTIVES:Chronic liver disease is a major cause of morbidity and mortality in the United States. Although often used to detect liver disease, the prevalence and etiology of elevated aminotransferases are unknown.METHODS:We analyzed data on adults ages 17 yr and older (n = 15,676) from the Third National Health and Nutrition Examination Survey (1988–1994). Participants were classified as having elevated aminotransferase levels if either aspartate aminotransferase or alanine aminotransferase was elevated above normal. Aminotransferase elevation was classified as “explained” if there was laboratory evidence of hepatitis B or C infection, iron overload, or if there was a history of alcohol consumption. Analyses were weighted to provide national estimates.RESULTS:The prevalence of aminotransferase elevation in the United States was 7.9%. Aminotransferase elevation was more common in men compared to women (9.3% vs 6.6%, p = 0.002), in Mexican Americans (14.9%) and non-Hispanic blacks (8.1%) compared to non-Hispanic whites (7.1%, p < 0.001). High alcohol consumption, hepatitis B or C infection and high transferrin saturation were found in only 31.0% of cases. Aminotransferase elevation was unexplained in the majority (69.0%). In both men and women, unexplained aminotransferase elevation was significantly associated with higher body mass index, waist circumference, triglycerides, fasting insulin, and lower HDL; and with type 2 diabetes and hypertension in women (all p < 0.05).CONCLUSIONS:Aminotransferase elevation was common in the United States, and the majority could not be unexplained by alcohol consumption, viral hepatitis or hemochromatosis. Unexplained aminotransferase elevation was strongly associated with adiposity and other features of the metabolic syndrome, and thus may represent nonalcoholic fatty liver disease.

Effect of an Internet-based Curriculum on Postgraduate Education: A Multicenter Intervention

We hypothesized that the Internet could be used to disseminate and evaluate a curriculum in ambulatory care, and that internal medicine residency program directors would value features made possible by online dissemination. An Internet-based ambulatory care curriculum was developed and marketed to internal medicine residency program directors. Utilization and knowledge outcomes were tracked by the website; opinions of program directors were measured by paper surveys. Twenty-four programs enrolled with the online curriculum. The curriculum was rated favorably by all programs, test scores on curricular content improved significantly, and program directors rated highly features made possible by an Internet-based curriculum.

Primary prevention of type 2 diabetes: integrative public health and primary care opportunities, challenges and strategies.

Type 2 diabetes imposes a large and growing burden on the publics health. This burden, combined with the growing evidence for primary prevention from randomized controlled trials of structured lifestyle programs leads to recommendations to include caloric reduction, increased physical activity and specific assistance to patients in problem solving to achieve modest weight loss as well as pharmacotherapy. These recommendations demand exploration of new ways to implement such primary prevention strategies through more integrated community organization, medical practice and policy. The US experience with control of tobacco use and high blood pressure offers valuable lessons for policy, such as taxation on products, and for practice in a variety of settings, such as coordination of referrals for lifestyle supports. We acknowledge also some notable exceptions to their generalizability. This paper presents possible actions proposed by an expert panel, summarized in Table 1 as recommendations for immediate action, strategic action and research. The collaboration of primary care and public health systems will be required to make many of these recommendations a reality. This paper also provides information on the progress made in recent years by the Division of Diabetes Translation at the US Centers for Disease Control and Prevention (CDC) to implement or facilitate such integration of primary care and public health for primary prevention.

The Challenge of Obesity-related Chronic Diseases

T ype 2 diabetes mellitus, hypertension, and hypercholesterolemia are common chronic diseases in the United States that are extremely costly to our society in terms of both health care expenditures and premature morbidity and mortality. Because all three diseases are strongly influenced by adiposity, clinical guidelines for each, as well as for obesity, recommend lifestyle modification as first-line management. 1–4 However, physicians do not routinely counsel their patients about lifestyle modification. 5–7

The use of clinical guidelines highlights ongoing educational gaps in physicians’ knowledge and decision making related to diabetes

BackgroundClinical guidelines for type 2 diabetes are a resource for providers to manage their patients and may help highlight specific areas in need of further education and training. We sought to determine how often guidelines are used and the relationship to physicians’ diabetes-related knowledge and decision making.MethodsExisting users of electronic clinical support tools were invited to complete an online questionnaire. A knowledge score was calculated for five questions related to prevention of diabetes and treatment of its complications. We explored the association of clinical guideline use with diabetes-related knowledge and self-reported decision making using logistic regression models, adjusted for key covariates.ResultsOf 383 physicians completing the questionnaire, 53% reported using diabetes guidelines routinely. Mean diabetes knowledge score for guideline users (GU) was significantly higher than non-guideline users (NGU) (3.37u2009±u20090.072 vs. 2.76u2009±u20090.084; pu2009<u20090.001). GU were significantly more likely to report a good understanding of type 2 diabetes medications (ORu2009=u20092.99, 95% CI 1.95-4.61; pu2009<u20090.001). GU were less likely to report their unfamiliarity with insulin as an important barrier to early insulin use (ORu2009=u20090.41, 0.21-0.80; pu2009=u20090.007) and with pharmacologic options as a barrier to prescribing intensive multifactorial interventions (ORu2009=u20090.32, 0.17-0.58; pu2009<u20090.001). Associations remained significant after adjusting for physician specialty, practice volume and frequency diagnosing or treating diabetes patients.ConclusionsSignificant gaps exist in diabetes-related knowledge and decision making among practicing physicians, as highlighted by clinical guideline use. The development of educational and training strategies to address these needs may ultimately improve outcomes for patients with diabetes and should be investigated in the future.

Baseline Predictors of Missed Visits in the Look AHEAD Study

To identify baseline attributes associated with consecutively missed data collection visits during the first 48 months of Look AHEAD—a randomized, controlled trial in 5,145 overweight/obese adults with type 2 diabetes designed to determine the long‐term health benefits of weight loss achieved by lifestyle change.