Atrial fibrillation (AF) is prevalent in patients with type 2 diabetes mellitus (DM) and is associated with markers of poor glycemic control; however, the impact of glycemic control on incident AF and outcomes is unknown. The aims of this study were to prospectively evaluate if intensive glycemic control in patients with DM affects incident AF and to evaluate morbidity and mortality in patients with DM and incident AF. A total of 10,082 patients with DM from the Action to Control Cardiovascular Risk in Diabetes (ACCORD) cohort were studied in a randomized, double-blind fashion. Participants were randomized to an intensive therapeutic strategy targeting a glycated hemoglobin level of <6.0% or a standard strategy targeting a glycated hemoglobin level of 7.0% to 7.9%. Incident AF occurred in 159 patients (1.58%) over the follow-up period, at a rate of 5.9 per 1,000 patient-years in the intensive-therapy group and a rate of 6.37 per 1,000 patient-years in the standard-therapy group (p = 0.52). In a multivariate model, predictors of incident AF were age, weight, diastolic blood pressure, heart rate, and heart failure history. Patients with DM and new-onset AF had a hazard ratio of 2.65 for all-cause mortality (95% confidence interval 1.8 to 3.86, p <0.0001), a hazard ratio of 2.1 for myocardial infarction (95% confidence interval 1.33 to 3.31, p = 0.0015), and a hazard ratio of 3.80 for the development of heart failure (95% confidence interval 2.48 to 5.84, p <0.0001). In conclusion, intensive glycemic control did not affect the rate of new-onset AF. Patients with DM and incident AF had an increased risk for morbidity and mortality compared with those without AF.
Highlights
- •
Intensive glycemic control did not alter the rate of incident AF.
- •
The investigators describe a number of predictors of AF in patients with diabetes.
- •
Incident AF is associated with adverse outcomes in patients with diabetes.
Type 2 diabetes mellitus (DM) confers roughly a 40% increased risk for developing atrial fibrillation (AF) compared with patients without DM. Excess cardiovascular morbidity and mortality are attributable to AF in patients with DM. In multiple studies, baseline AF was associated with a significantly greater risk for all-cause mortality and comparable higher risk for cardiovascular death, stroke, and heart failure ; incident AF in patients with DM leads to increased risk for stroke and heart failure. Risk for AF in patients with DM has been linked to glycemic control. In treated patients with DM with new AF in a population-based case-control study, every 1% increase in glycated hemoglobin (HbA 1c ) increased the risk for AF by 14%. In a large study, HbA 1c and fasting glucose levels were positively and independently correlated with incident AF in patients with DM. Prospectively, a series of small clinical trials have demonstrated reduced incidence of perioperative AF in patients with DM who underwent coronary artery bypass grafting subjected to more stringent glycemic control. Whether long-term glycemic control alters incident AF or outcomes in patients with DM is unknown. The Action to Control Cardiovascular Risk in Diabetes (ACCORD) trial was designed to address the impact of intensive glycemic control, compared with standard treatment, on cardiovascular outcomes. Those in the intensive-therapy arm had increased mortality and complication rates. In the present analysis, we examined the impact of intensive versus standard glycemic control on the incidence of AF and evaluated the impact of AF on mortality and other cardiovascular events in patients with incident AF in the ACCORD trial.
Methods
The rationale and design of the ACCORD trial have been previously described. Briefly, this was a National Heart, Lung, and Blood Institute–sponsored trial conducted at 77 clinical centers across the United States and Canada. Volunteers with DM, HbA 1c levels ≥7.5, and ages of 40 to 79 years were recruited. By inclusion criteria, subjects either had cardiovascular disease or were aged of 55 to 79 years and had anatomic evidence of significant atherosclerosis, albuminuria, left ventricular hypertrophy, or ≥2 additional cardiovascular risk factors (dyslipidemia, hypertension, current smoking status, or obesity). Those with either frequent or recent hypoglycemic events, body mass indexes >45 kg/m 2 , serum creatinine levels >1.5 mg/dl, serious illnesses, or unwillingness to participate in home glucose monitoring or insulin injection were excluded.
All 10,251 patients were randomly assigned to receive intensive therapy targeting an HbA 1c level of <6.0% or to receive standard therapy targeting an HbA 1c level of 7.0% to 7.9%. Regimens were individualized at the investigators’ and patients’ discretion, with careful auditing of adverse responses to therapies. Patients in the intensive-therapy group were followed monthly in the clinic for the first 4 months, then every 2 months thereafter with additional visits as necessary; patients in the standard-therapy group attended visits every 4 months. The intensive glycemia intervention was stopped early because of increased mortality in the glycemia control arm, and all participants were transitioned to the standard glycemia control intervention.
Twelve-lead electrocardiograms were obtained at the start of the trial and at the biennial follow-up visits (i.e., every 2 years) and the close-out visit. At all ACCORD sites, electrocardiograms were digitally acquired using a GE MAC 1200 electrocardiograph (GE Healthcare, Milwaukee, Wisconsin) at 10 mm/mV calibration and a speed of 25 mm/s. Reading of electrocardiograms was performed centrally at the Epidemiological Cardiology Research Center at Wake Forest School of Medicine (Winston-Salem, North Carolina). All electrocardiograms were initially inspected visually for technical errors and inadequate quality before being automatically processed using a GE 12-SL Marquette version 2001 (GE Healthcare). Electrocardiographic abnormalities, including AF were classified and coded using the Minnesota ECG Classification.
Adverse events included in our study were the prespecified primary outcome (a composite of the first occurrence of nonfatal myocardial infarction (MI) or nonfatal stroke or death from cardiovascular causes), any MI, nonfatal MI, death from any cause, death from a cardiovascular cause, any stroke, nonfatal stroke, congestive heart failure, angina, or revascularization. Cardiovascular death included death from an MI, congestive heart failure, arrhythmia, invasive cardiovascular intervention, stroke, sudden cardiac death, and death from other vascular disease.
Baseline characteristics were compared in the 2 study groups using chi-square tests and 2-sample Student’s t tests. Analyses of outcomes were performed using time-to-event methods according to the intention-to-treat principle, and occurrences of these outcomes in the 2 study groups were compared using hazard ratios and 95% confidence intervals. Two-sided p values were obtained from likelihood ratio tests from Cox proportional-hazards regression analyses. New-onset AF rates are expressed as events per 1,000 patient-years, taking into account censoring of follow-up data. Kaplan-Meier estimates were used to obtain the proportion of patients who developed AF during follow-up. The risks of events associated with the incidence of AF were estimated using Cox proportional-hazards models, with adjustment for potential confounding baseline covariates that were thought to be of prognostic importance for the outcomes of interest. All statistical analyses were conducted at the coordinating center using S-Plus version 8.0 (Insightful, Seattle, Washington) or SAS software, version 9.1 (SAS Institute, Cary, North Carolina).
Results
Of the 10,251 subjects enrolled in ACCORD, 10,209 had baseline electrocardiographic data available, and of these patients, 10,082 without baseline AF were included in our analysis. Incident AF was diagnosed in 159 of the 10,082 participants (1.58%) by electrocardiography at 1 of the follow-up visits over a median of 4.68 years. The mean age of the group was 62.2 ± 6.8 years, with approximately 39% women. The average HbA 1c level at baseline was 8.3%. The median HbA 1c level achieved at 1 year, and maintained throughout the follow-up period, was 6.4% (interquartile range 6.1% to 7.0%) and 7.5% (interquartile range 7.0% to 8.1%), respectively. Baseline characteristics of the study groups were similar and are listed in Table 1 .
| Characteristic | Standard (N = 5042) | Intensive (N = 5040) | p – Value |
|---|---|---|---|
| Age (years) | 62.2 ± 6.8 | 62.2 ± 6.8 | 0.94 |
| Female sex | 38.6% | 38.8% | 0.87 |
| White | 64.9% | 64.1% | 0.40 |
| Black | 19.0% | 19.9% | 0.27 |
| Hispanics | 7.5% | 7.0% | 0.33 |
| Weight (kgs) | 93.5 ± 15.6 | 93.4 ± 18.6 | 0.86 |
| Waist (cm) | 106.7 ± 13.7 | 106.7 ± 14.0 | 0.81 |
| Body mass index (kg/m 2 ) | 32.2 ± 5.5 | 32.2 ± 5.5 | 0.74 |
| Smoking | 13.7% | 14.3% | 0.45 |
| Duration of DM (years) | 10.9 ± 7.8 | 10.8 ± 7.8 | 0.58 |
| ECG LV hypertrophy | 5.2% | 5.2% | 0.99 |
| Systolic blood pressure (mmHg) | 136.5 ± 17.2 | 136.3 ± 17.0 | 0.49 |
| Diastolic blood pressure (mmHg) | 75.0 ± 0.6 | 74.8 ± 10.6 | 0.37 |
| Pulse pressure (mmHg) | 61.5 ± 14.7 | 61.5 ± 14.6 | 0.87 |
| Heart rate (bpm) | 72.5 ± 11.7 | 72.7 ± 11.7 | 0.38 |
| Previous CV event | 34.9% | 35.5% | 0.56 |
| Previous heart failure | 4.6% | 4.6% | 0.97 |
| Medications | |||
| Insulin | 35.7% | 34.0% | 0.073 |
| Medtformin | 60.2% | 59.8% | 0.70 |
| Any sulfonylurea | 49.4% | 50.8% | 0.17 |
| Any antihypertensive agent | 85.9% | 84.7% | 0.10 |
| ACE-inhibitors | 53.0% | 52.8% | 0.86 |
| ARBs | 16.4% | 15.6% | 0.29 |
| Any thiazide diuretic | 19.1% | 19.5% | 0.64 |
| Beta-blockers | 29.7% | 28.4% | 0.13 |
| Alpha-beta blockers | 2.8% | 2.5% | 0.42 |
| Dihydropyridine CCBs | 11.7% | 11/6% | 0.82 |
| Non-dihydropyridine CCBs | 7.3% | 6.7% | 0.20 |
| Aspirin | 54.5% | 55.2% | 0.51 |
| Statin | 62.5% | 61.7% | 0.46 |
| Fibrate | 5.7% | 6.1% | 0.40 |
| Fasting serum glucose (mg/dl) | 175.6 ± 56.4 | 174.9 ± 56.0 | 0.58 |
| Glycated hemoglobin (%) | 8.3 ± 1.1 | 8.3 ± 1.1 | 0.50 |
| Serum creatinine (mg/dl) | 0.91 ± 0.23 | 0.91 ± 0.24 | 0.20 |
| Urine albumin/creatinine ratio (mg/g) | 93.5 ± 315.8 | 97.3 ± 394.0 | 0.59 |
| Potassium (mmol/L) | 4.48 ± 0.65 | 4.48 ± 0.44 | 0.96 |
| Total cholesterol (mg/dl) | |||
| LDL cholesterol | 105.0 ± 33.9 | 105.0 ± 33.9 | 0.97 |
| HDL cholesterol | 42.0 ± 1.5 | 41.9 ± 11.8 | 0.65 |
| Triglycerides | 189.5 ± 149.0 | 191.0 ± 148.8 | 0.60 |
New-onset AF occurred at a rate of 5.9 per 1,000 patient-years in the intensive group and a rate of 6.37 per 1,000 patient-years in the standard group (p = 0.52). Univariate predictors of incident AF are listed in Table 2 . After stepwise regression analysis, remaining multivariate predictors were age, weight, diastolic blood pressure, heart rate, and previous heart failure, with hazard ratios listed in Table 3 . Among the entire cohort, AF was associated with a higher incidence of the primary outcome: first occurrence of nonfatal MI or nonfatal stroke or death from cardiovascular cause, all-cause mortality, cardiovascular death, any MI, nonfatal MI, and congestive heart failure. When stratified by study group, AF did not reach a significant association with MI or nonfatal MI in the standard-therapy group; however, this association remained in the intensive-therapy group. AF was not associated with stroke, nonfatal stroke, angina, or revascularization ( Figure 1 ).
| Characteristic | Never had AF (N = 9,958) | Developed New AF (N = 159) | p – Value |
|---|---|---|---|
| Age (years) | 62.1 ± 5.8 | 66.4 ± 6.3 | <0.0001 |
| Female sex | 38.9% | 27.7% | 0.004 |
| White | 64.2% | 84.9% | <0.0001 |
| Black | 19.7% | 6.9% | <0.0001 |
| Hispanics | 19.7% | 1.9% | <0.0001 |
| Weight (kgs) | 93.3 ± 16.5 | 100.4 ± 20.1 | <0.0001 |
| Waist (cm) | 106.6 ± 13.8 | 111.2 ± 14.9 | <0.0001 |
| Body mass index (kg/m 2 ) | 32.2 ± 5.5 | 33.4 ± 5.5 | 0.007 |
| Smoking | 27.9% | 25.8% | 0.70 |
| Duration of DM (years) | 10.87 ± 7.77 | 10.37 ± 7.67 | 0.42 |
| ECG LV hypertrophy | 5.2% | 4.4% | 0.66 |
| Systolic blood pressure (mmHg) | 136.4 ± 17.1 | 137.4 ± 16.6 | 0.45 |
| Diastolic blood pressure (mmHg) | 74.9 ± 10.6 | 71.0 ± 10.1 | <0.0001 |
| Pulse pressure (mmHg) | 61.4 ± 14.6 | 66.4 ± 16.4 | 0.0002 |
| Heart rate (bpm) | 72.7 ± 11.7 | 67.2 ± 11.8 | <0.0001 |
| Previous CV event | 35.0% | 41.5% | 0.090 |
| Previous heart failure | 4.5% | 10.1% | 0.021 |
| Medications | |||
| Insulin | 34.9% | 34.0% | 0.81 |
| Medtformin | 60% | 54.6% | 0.38 |
| Any sulfonylurea | 50.0% | 47.8% | 0.55 |
| Any antihypertensive agent | 85.2% | 93.7% | <0.0001 |
| ACE-inhibitors | 52.8% | 58.5% | 0.15 |
| ARBs | 16.0% | 18.2% | 0.44 |
| Any thiazide diuretic | 19.3% | 18.2% | 0.73 |
| Beta-blockers | 28.8% | 42.1% | 0.0002 |
| Alpha-beta blockers | 2.6% | 3.8% | 0.45 |
| Dihydropyridine CCBs | 11.6% | 13.8% | 0.39 |
| Non-dihydropyridine CCBs | 7.0% | 10.7% | 0.13 |
| Aspirin | 54.7% | 62.2% | 0.056 |
| Statin | 62.1% | 63.5% | 0.71 |
| Fibrate | 0.59% | 8.2% | 0.30 |
| Fasting serum glucose (mg/dl) | 175.2 ± 56.3 | 175.4 ± 48.8 | 0.96 |
| Glycated hemoglobin (%) | 8.30 ± 1.05 | 8.31 ± 0.95 | 0.92 |
| Serum creatinine (mg/dl) | 0.909 ± 0.230 | 1.003 ± 0.264 | <0.0001 |
| Urine albumin/creatinine ratio (mg/g) | 95.8 ± 359.2 | 95.5 ± 230.2 | 0.99 |
| Potassium (mmol/L) | 4.48 ± 0.56 | 4.51 ± 0.46 | 0.31 |
| Total cholesterol (mg/dl) | |||
| LDL cholesterol | 105.0 ± 33.9 | 101.9 ± 30.6 | 0.26 |
| HDL cholesterol | 41.9 ± 11.7 | 41.6 ± 9.8 | 0.68 |
| Triglycerides | 190.4 ± 149.3 | 174.5 ± 103.0 | 0.059 |
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
