Author + information
- Received December 12, 2012
- Revision received February 6, 2013
- Accepted February 26, 2013
- Published online August 6, 2013.
- Raffaele Marfella, MD, PhD∗∗ (, )
- Ferdinando Carlo Sasso, MD, PhD†,
- Mario Siniscalchi, MD, PhD‡,
- Mario Cirillo, MD§,
- Pasquale Paolisso∗,
- Celestino Sardu, MD∗,
- Michelangela Barbieri, MD, PhD∗,
- Maria Rosaria Rizzo, MD, PhD∗,
- Ciro Mauro, MD† and
- Giuseppe Paolisso, MD, PhD∗
- ∗Department of Geriatrics and Metabolic Diseases, Second University of Naples, Naples, Italy
- †Department of Internal and Experimental Medicine, Center of Cardiovascular Excellence, Second University of Naples, Naples, Italy
- ‡Department of Cardiology, Hospital Cardarelli, Naples, Italy
- §Department of Neuroradiology, Second University of Naples, Naples, Italy
- ↵∗Reprint requests and correspondence:
Dr. Raffaele Marfella, Geriatrics and Metabolic Disease, Second University of Naples, Piazza Miraglia, 2, 80138 Naples, Italy.
Objectives This study evaluated whether subclinical episodes of atrial fibrillation (AF) were associated with an increased risk of silent cerebral infarct (SCI) and stroke in diabetic patients younger than 60 years who did not have other clinical evidence of AF and cerebrovascular disease at baseline.
Background In type 2 diabetic patients, one-fourth of strokes are of unknown cause, and subclinical episodes of AF may be a common etiologic factor.
Methods A total of 464 type 2 diabetic patients younger than 60 years were included in a longitudinal observational study and matched to patients without diabetes. Patients underwent 48-h electrocardiographic Holter monitoring quarterly to detect brief subclinical episodes of AF (duration of AF <48 h) and were followed up for 37 months. The outcomes were SCI, assessed by magnetic resonance imaging of the brain, and stroke events during the follow-up period.
Results The prevalence of subclinical episodes of AF was significantly greater among patients with diabetes compared with matched healthy subjects (11% vs. 1.6%, p < 0.0001). During an average duration of 37 months, 43 stroke events occurred in the diabetic population and no events occurred in healthy subjects. Diabetic patients with silent episodes of AF (n = 176) had a higher baseline prevalence of SCI (61% vs. 29%; p < 0.01) and a higher number of stroke events (17.3% vs. 5.9%; p < 0.01) during the follow-up period than the other patients (n = 288). An episode of silent AF was an independent determinant of SCI (odds ratio: 4.441; p < 0.001; confidence interval: 2.42 to 8.16) and an independent predictor of the occurrence of stroke in diabetic patients (hazard ratio: 4.6; p < 0.01; confidence interval: 2.7 to 9.1).
Conclusions Subclinical episodes of AF occurred frequently in type 2 diabetic patients and were associated with a significantly increased risk of SCI and stroke.
The relative risk of stroke in diabetic patients younger than 60 years is twice that of nondiabetic subjects older than 70 years (1). Diabetes-related factors (e.g., hyperglycemia, insulin resistance), diabetes-associated vascular risk factors (e.g., hypertension and dyslipidemia), and genetic, demographic, and lifestyle factors may modulate such risk (2). The contribution of these factors, many of which are strongly interrelated, is likely to differ according to the type of diabetes and the patient's age. Nevertheless, after adjustment for the previously mentioned risk factors, patients with diabetes have a doubled risk of stroke compared with those without diabetes (1). Moreover, only 15% of diabetic patients with stroke have a history of symptomatic atrial fibrillation (AF) (3). Thus, approximately one-fourth of strokes are of unknown cause, and subclinical AF may be a common etiologic factor (4). The risk of AF, a major cause of thromboembolic stroke, is increased by 40% in diabetic patients (5). Permanent AF, end stage of progression from paroxysmal forms (6), and silent paroxysmal AF in patients without previously documented arrhythmic episodes are associated with at least the same risk of stroke as symptomatic AF (7). Diabetes is an independent determinant of AF, and diabetic patients frequently have asymptomatic (e.g., silent) AF (8,9), but it is unknown whether silent episodes of AF influence cerebrovascular diseases in diabetic patients. In this context, the identification of silent episodes of AF as a possible cause of excess stroke in young/adult diabetic patients may have an important role in clinical and therapeutic management. Consequently, there is a need for more evaluations addressing the pathophysiological alterations that underlie the cerebrovascular diseases and silent episodes of AF in diabetic patients. Our study prospectively evaluated whether asymptomatic episodes of AF are more closely related to silent cerebral infarct (SCI) and stroke events in type 2 diabetic patients who did not have other evidence of AF or stroke at baseline.
The study protocol included a recruitment period of 4 years (from January 1, 2005, to January 1, 2009) and a follow-up period of 3 years. In the present study, 464 type 2 diabetic patients (mean age, 52 ± 6 years) were followed up prospectively for an average of 37 months. These patients were selected from a larger cohort of type 2 diabetic patients (N = 1,992) from 4 participating institutions (1 clinic, 2 hospitals, and 1 outpatient clinic). The inclusion criteria were age younger than 60 years, successful quarterly 48-h electrocardiographic (ECG) Holter monitoring (48HM), and assessment of the presence/absence of SCI by magnetic resonance imaging (MRI) of the brain. Patients with any of the following criteria at initial diagnosis were excluded: arrhythmia, documented persistent and/or permanent AF, documented stroke or transient ischemic attack, anticoagulation therapy, coronary artery or valvular heart disease, cardiomyopathy, history of congestive heart failure, hypertension, carotid and peripheral vascular disease, hyperthyroidism, chronic obstructive pulmonary disease, obstructive sleep apnea, and hepatic damage. As a result, 1,528 of the 1,992 screened diabetic patients were excluded and 464 were included in the study. All selected diabetic patients were matched to 240 healthy subjects without diabetes (control group) using anthropometric data. The control group was recruited at our institution from a study population of healthy subjects without exclusion criteria. These subjects were managed to provide anthropometric characteristics comparable to the other groups. All patients studied were ambulatory and gave informed consent to participate in the study, which was approved by the Ethics Committee for Human Studies at Second University of Naples (Naples, Italy).
All selected patients underwent 48-h ambulatory ECG recording (EVO, Spacelabs Healthcare, Hertford, United Kingdom) at 3, 6, 9, and 12 months during the screening for AF and then annually for another 3 years. The system was fully automatic and computerized; tracings were analyzed by 2 investigators who were unaware of the results of other investigations. Three ECG leads (modified leads V1, V5, and aVF) and a time signal to correct for tape speed irregularities were recorded. The quarterly 48HM follow-up was performed in all patients to identify silent episodes of AF (10). Baseline AF was classified as episode, paroxysmal, persistent, or permanent. Silent (i.e., asymptomatic) AF may present as any of the temporal forms of AF (11). An episode of AF was defined as a duration of AF <48 h. The duration of an episode of AF was arbitrarily classified as follows: 10 min to 1 h, 1 h to 24 h, and >24 to 48 h (10). The average duration (hours) of episodes of AF over the whole observational period was defined as the absolute AF burden. We classified the patients according to the presence of silent episodes of AF (SAFE group, n = 176) and the absence of silent episodes of AF (non-SAFE group, n = 288). In all patients, MRI of the brain was performed with a superconducting magnet with a main strength of 1.5-T within 30 days of the baseline 48HM (12). T1- and T2-weighted images were obtained in the transverse plane with 7.8-mm-thick or 8.0-mm-thick sections. An SCI was defined as a low signal intensity area (3 to 15 mm) on T1-weighted images that was also visible as a hyperintense lesion on T2-weighted images. All SCIs detected were lacunar infarcts with a size of <15 mm. The MRIs of the subjects were randomly stored and interpreted without knowledge of the subjects' names and characteristics.
Post-screening follow-up of the patients was scheduled annually for 36 months at our outpatient clinic or at the referring cardiologists' office. During these visits, 48HM rhythm documentation and documentation of clinical data as well as complications were performed. All patients were treated with aspirin 75 to 325 mg/day; if the CHADS2 (Congestive heart failure, Hypertension, Age ≥75 years, Diabetes mellitus, previous Stroke/transient ischemic attack) score was ≥2, antiaggregant therapy was switched to oral anticoagulation therapy (13). Stroke events were diagnosed by the physician who was caring for the patient at the time of the event, and independent neurologists reviewed the cases and confirmed the diagnosis of stroke events. Stroke was diagnosed on the basis of sudden onset of a neurological deficit that persisted for >24 h in the absence of any other disease process that could explain the symptom.
Data are expressed as mean ± SD. Two-sided unpaired t test and chi-square test were used to test differences between the groups in the mean values of continuous measures and prevalence rates, respectively. Hazard ratios (HRs) and odds ratios (ORs) with 95% confidence intervals (CIs) were calculated with Cox proportional hazards regression models and multiple logistic regression analysis, respectively. In particular, Cox regression multivariate analysis, including sex, body mass index, diastolic blood pressure, duration of diabetes, plasma glucose level, hemoglobin A1c, interventricular septum and hyperlipidemia, silent episodes of AF, systolic blood pressure, and prevalence of SCI as covariates, was performed to identify the independent prognostic factors for predicting stroke. We used Kaplan-Meier survival analysis to determine the probability that ictus would occur in the SAFE and non-SAFE groups, comparing the 2 groups with the use of the log-rank test. The statistical calculations were performed using SPSS version 12.0 (SPSS Inc., Chicago, Illinois). Differences with p < 0.05 were considered to be statistically significant.
Cumulative quarterly 48HM (192 h for each patient) showed a greater prevalence of subclinical episodes of AF among patients with diabetes (n = 1,992) compared with matched healthy subjects (n = 240) (11% [n = 212] vs. 1.6% [n = 4]; p < 0.001). Of the 1,992 screened patients with type 2 diabetes, 176 (9%) met the inclusion criteria for silent episodes of AF and clinical characteristics (SAFE group), whereas 288 (15%) met the inclusion criteria for only clinical characteristics (non-SAFE group). Thus, the present study included 464 type 2 diabetic patients and 240 healthy subjects. Table 1 shows the baseline characteristics of the study populations. No significant differences in clinical and anthropometric characteristics were found between the study groups. Aspirin was used by 99% of the diabetic patients in both the SAFE and non-SAFE groups and 10% of the subjects in the control group, and none of the patients were being treated with a vitamin K antagonist at baseline. The quarterly 48HM follow-up revealed 12 episodes of silent AF lasting 10 min to 1 h (39 ± 10 min), 151 episodes lasting 1 to 24 h (12 ± 6 h), and 51 episodes lasting ≥24 to <48 h (37 ± 6 h) in diabetic patients in the SAFE group, whereas only 2 episodes of AF lasting 1 to 24 h were revealed in healthy subjects. During cumulative quarterly 48HM, the mean absolute burden of AF was 21 ± 15 h in the 176 patients in the SAFE group and 3 ± 1.4 h in the 2 healthy subjects with episodes of AF (p < 0.001). SCI was detected in 190 diabetic patients (41%; n = 464) and only 1 healthy subject (0.5%). On MRI examination, SCI was more frequently detected in the SAFE group than in the non-SAFE group (61% vs. 29%; p < 0.01). Moreover, the absolute burden of AF was significantly correlated with both size (r = 0.574; p < 0.001) and number of SCIs (r = 0.591; p < 0.001) (Fig. 1). Multiple logistic regression analysis demonstrated that episodes of silent AF were associated with SCI (p = 0.001) as well as left atrial diameter (p = 0.024), systolic blood pressure (p = 0.002), and duration of diabetes (p = 0.044), whereas other confounders (sex, body mass index, diastolic blood pressure, plasma glucose level, hemoglobin A1c, interventricular septum, and hyperlipidemia) were again not significant (Table 2).
After a mean follow-up period of 37.5 ± 1.6 months, clinical AF developed in 20 patients (11%) with silent episodes of AF compared with only 13 patients (4%) without silent episodes of AF (p < 0.01) (Table 1). All patients with clinical AF were excluded from follow-up analysis. No AF was diagnosed in healthy subjects. Therefore, the stroke follow-up was performed in 156 patients in the SAFE group and 275 patients in the non-SAFE group. Over the course of the follow-up period, 26 patients in the SAFE group (15%) and 19 (7%) in the non-SAFE group were treated with a vitamin K antagonist, including 33 patients who had developed clinical AF. Despite the use of antiplatelet agents and a vitamin K antagonist, 43 stroke events occurred during follow-up. As shown in Figure 2, the cumulative percent of stroke-free survival was significantly lower in patients in the SAFE group. After 37.5 ± 1.6 months, ischemic stroke developed in 27 patients in the SAFE group (17.3%) compared with 16 patients in the non-SAFE group (5.9%) (p < 0.01). No stroke occurred in healthy subjects. Interim analysis revealed that the annual incidence of stroke was higher in patients in the SAFE group compared with patients in the non-SAFE group during the first (3.8% vs. 1.4%; p < 0.05) and second year (6.4% vs. 2.2%; p < 0.05) of follow-up (Fig. 2). Cox regression analysis revealed that the presence of silent episodes of AF (HR: 4.6; 95% CI: 2.7 to 9.1; p < 0.001), systolic blood pressure (HR: 1.7; 95% CI: 1.02 to 2.92; p < 0.01), and prevalence of SCI (HR: 3.1; 95% CI: 1.3 to 7.1; p < 0.005) were associated with risk of stroke, whereas there were no significant other confounders (sex, body mass index, diastolic blood pressure, duration of diabetes, plasma glucose level, hemoglobin A1c, interventricular septum, and hyperlipidemia). Of the 43 stroke events, 42 were ischemic and 1 was hemorrhagic.
The first relevant finding of this study was that type 2 diabetic patients younger than 60 years and without clinical AF had a significant incidence of subclinical brief episodes of AF. Indeed, cumulative quarterly 48HM showed a significantly greater prevalence of subclinical episodes of AF among patients with diabetes compared with matched healthy subjects. Subclinical brief episodes of AF were detected in approximately one of the hundreds of healthy subjects within 12 months after quarterly 48HM evaluations and, in a more marked fashion, in one-tenth of the diabetic patients.
The second major finding of the study was that the presence of brief episodes of AF was associated with an increased risk of SCI as well as the subsequent development of stroke in type 2 diabetic patients younger than 60 years of age independently of duration of diabetes and target organ damage. The prevalence of SCI was 41% among the diabetic patients, which is comparable to previous observations (approximately 42%) (14). As background for such an association, the brief episodes of AF seem to have a pivotal role. It seems likely that SCI exists more frequently in type 2 diabetic patients with silent episodes of AF than in subjects without it, because 61% of patients in the SAFE group had SCI compared with 29% in the non-SAFE group. Intriguingly, because the OR in our population was 4.41, the results of our study suggest that silent episodes of AF were an independent determinant of SCI. Moreover, the risk of SCI was higher when subclinical episodes of AF were of longer duration, as evidenced by a positive correlation between AF burden and SCI. Notably, our data are consistent with previous observations that SCI is associated with an increased risk of subsequent stroke in patients with AF (15). However, previous studies did not evaluate the role of brief episodes of AF in cerebrovascular diseases in type 2 diabetic patients. In our study, the baseline prevalence of SCI was associated with stroke risk in patients in the SAFE group. According to our data, the presence of SCI increased the risk of subsequent stroke by 2 to 4 times in the general population, independently of cardiovascular risk factors (16,17). Furthermore, SCI is reported to be a strong predictor of subsequent clinical stroke (approximate OR: 6 to 10) (18,19) and can be considered the most important diabetic target organ damage marker for stroke. In light of the data presented here, silent brain infarcts should not be considered just intermediates in the relationship between vascular risk factors and the risk of stroke but markers for other factors, such as brief episodes of AF, that lead to stroke.
During a mean follow-up period of 37 months, stroke developed in approximately 9% of the diabetic patients while they were receiving therapy with antiplatelet agents. The SAFE group had a higher incidence of stroke (15%) than the non-SAFE group (5%). Furthermore, the annual risk of stroke was even higher (3.5% to 6.8%) in diabetic patients in the SAFE group compared with diabetic patients in the non-SAFE group. This incidence is 2-fold higher than that seen in diabetic patients without silent episodes of AF (1.4% to 2.1%) and comparable to the previously observed incidence of stroke (approximately 8%) in patients with type 2 diabetes and cardiovascular complications (2). In the present study, brief episodes of AF were cross-sectionally significantly associated with the baseline prevalence of SCI detected by MRI of the brain and prospectively with stroke events in type 2 diabetic patients who did not have clinical AF.
Despite the relatively large number of asymptomatic episodes of AF documented in this prospective study by quarterly 48HM, the study could be underpowered for the technical approach used. Indeed, even if standardized recording technique were used, it is possible that many silent episodes of AF were missed. This could explain the relatively high risk of cerebrovascular events in those patients without documented AF. Strong evidence suggests that a commonly used intermittent follow-up strategy (24-h electrocardiographic Holter monitoring) is significantly inferior to full disclosure heart rhythm observation by implantable continuous cardiac rhythm monitoring with respect to detection of AF recurrence (20). However, quarterly 48HM may improve the heart rhythm investigations. In addition, we could not use continuous heart rhythm surveillance in our population because there was no indication for a device implanted subcutaneously. Finally, on the basis of our data, we cannot establish a temporal link between the episodes of AF and cerebrovascular disease. The SCIs are probably due to similar AF that occurred in patients with silent AF before this study.
However, further research in a larger sample of populations is needed to confirm the reproducibility of our new findings. The identification of “brief episodes of AF” in type 2 diabetic patients may have clinical relevance in the identification of patients at risk and in the implementation of preventive measures for stroke, even after assessment of target organ damage.
The authors have reported that they have no relationships relevant to the content of this paper to disclose.
- Abbreviations and Acronyms
- atrial fibrillation
- confidence interval
- 48-h electrocardiographic Holter monitoring
- hazard ratio
- magnetic resonance imaging
- odds ratio
- silent cerebral infarct
- Received December 12, 2012.
- Revision received February 6, 2013.
- Accepted February 26, 2013.
- American College of Cardiology Foundation
- Kissela B.M.,
- Khoury J.,
- Kleindorfer D.,
- et al.
- Calkins H.,
- Kuck K.H.,
- Cappato R.,
- et al.
- Nichols G.A.,
- Reinier K.,
- Chugh S.S.
- Hanke T.,
- Charitos E.I.,
- Stierle U.,
- et al.
- Fuster V.,
- Rydén L.E.,
- Cannom D.S.,
- et al.
- You J.J.,
- Singer D.E.,
- Howard P.A.,
- et al.,
- American College of Chest Physicians
- Nomura K.,
- Hamamoto Y.,
- Takahara S.,
- et al.
- Kempster P.A.,
- Gerraty R.P.,
- Gates P.C.
- Vermeer S.E.,
- Hollander M.,
- van Dijk E.J.,
- Hofman A.,
- Koudstaal P.J.,
- Breteler M.M.B.
- Kario K.,
- Shimada K.,
- Matsuo T.,
- et al.
- Kobayashi S.,
- Okada K.,
- Koide H.,
- et al.