Short Leukocyte Telomere Length Is Associated With Cardioembolic Stroke Risk in Patients With Atrial Fibrillation
Background and Purpose—The risk of cardioembolic stroke in patients with atrial fibrillation (AF) cannot be accurately assessed and novel tools are needed to improve prediction. We hypothesize that telomere shortening constitutes a novel risk factor for cardioembolic stroke in patients with AF.
Methods—The peripheral blood leukocyte telomere length (LTL) was determined by real-time polymerase chain reaction in 187 patients with AF, 93 of them without stroke history and 94 of them having suffered 1 cardioembolic stroke. Percentiles were calculated according to LTL values in the nonstroke group to estimate the cardioembolic stroke risk associated with LTL using logistic regression models.
Results—Short LTL values were independently and dose-dependently associated with an increased risk of cardioembolic stroke, with an odds ratio (95% confidence interval) of 2.93 (1.24–6.94) and 6.26 (2.01–19.52), respectively, for sex, hypertension, diabetes mellitus, heart failure, and age-adjusted models using the LTL 10th and 5th percentile cut-offs, respectively.
Conclusions—Telomere shortening is associated with cardioembolic stroke risk in patients with AF. Prospective studies are encouraged to establish the value of LTL to improve prediction tools to categorize cardioembolic stroke risk in AF.
Atrial fibrillation (AF) is the major risk factor for cardioembolic stroke.1 CHADS2 (congestive heart failure, hypertension, age≥75 years, diabetes mellitus, stroke/transient ischemic attack [doubled]) or CHA2DS2-VASc (congestive heart failure, hypertension, age≥75 years [doubled], diabetes mellitus, stroke/transient ischemic attack [doubled], vascular disease, age 65–74 years, and sex category [female]), used to risk stratify patients with AF, are not accurate enough and >2% of patients per year are considered to be at low risk experience of stroke.2 Thus, novel tools to improve risk stratification are needed.
Telomeres are tandem TTAGGG repeats at the chromosome end to avoid unwanted events, such as chromosomal fusion. Their shortening is a consequence of cell division, and it has the consequence that, once a threshold is exceeded, cells enter a senescent/apoptotic state. Environmental factors such as oxidative stress or inflammation also contribute to shortening.3 Telomere length of peripheral blood leukocytes telomere length (LTL) adequately reflects telomere length of other cells, such as vascular cells.3,4 Importantly, short LTL has been associated with a variety of cardiovascular diseases.4–12 However, its association with the cardioembolic subtype of stroke has not been addressed. Because cardioembolic stroke risk increases with age in AF1 and, in turn, age influences telomere length,3 we hypothesized that short LTL is associated with cardioembolic stroke risk in patients with AF.
A total of 204 patients with nonrheumatic AF, of whom 102 had never had a stroke and 102 had had 1 cardioembolic stroke, were consecutively recruited when attending the Anticoagulation Clinic of the University Hospital of Salamanca, Spain, between May 2009 and February 2010. The diagnosis, inclusion, and exclusion criteria are explained in the expanded Methods section of the online-only Data Supplement. The study was approved by the Institutional Review Boards of the Hospital of Salamanca and University of Navarra. All patients gave informed consent.
The genomic DNA of peripheral blood leukocytes was isolated to measure LTL by a real-time polymerase chain reaction–based method13 as described in the expanded Methods section of this article.
Continuous and categorical variables were compared between AF stroke and nonstroke patients using the Student’s t and the χ2 tests, respectively. Correlation was assessed using Spearman correlation coefficient.
A nonconditional logistic regression model was used to evaluate the LTL-associated stroke risk. We assessed the goodness of fit with the Hosmer–Lemeshow goodness-of-fit statistic. The main independent variable was LTL, of which the 10th and 5th percentiles according to the nonstroke group values were used as cut-offs. Univariate and sex, hypertension, diabetes mellitus, heart failure, and age-adjusted multivariate models were performed. Stata/SE 13.1 (StataCorp, College Station, TX) was used for the analyses.
Table 1 shows the patients' clinical characteristics. Seventeen patients were finally excluded because of lack of DNA. Patients with Stroke exhibited higher CHADS2 values. There were no differences in sex, age, international normalized ratio, or creatinine. Remarkably, LTL was significantly shorter in the patients with stroke (Table 1; Figure I in the online-only Data Supplement).
Because advanced age has been associated with telomere shortening,3 we studied their correlation in our population. We found no significant correlation between age and LTL (ρ=0.1; P=0.16). This unexpected result may be explained by the narrow range of age of our cohort.
To gain insights into the cardioembolic stroke risk associated with telomere shortening, nonconditional logistic regression models were applied, using the 10th and 5th percentiles of the nonstroke group LTL values as cut-offs. As shown in Table 2, LTL below the 10th percentile significantly increased the risk of cardioembolic stroke. Interestingly, the risk remained unchanged when sex and CHADS2 components were used as confounders. It was noteworthy that the risk became higher when the 5th percentile was applied. Thus, short LTL independently and dose-dependently increases the cardioembolic stroke risk.
New tools are needed to improve the assessment of cardioembolic stroke risk in patients with AF. We provide evidence that, among the patients with AF, cardioembolic stroke risk is dose-dependently associated with short LTL. Remarkably, the risk seems to be independent of traditional stroke risk factors.
The association between LTL and stroke has been a matter of controversy.8–12 However, those studies never specifically addressed the cardioembolic subtype: 2 reports excluded these patients,8,11 and the others included only a small proportion of them within the study population.9,10,12 Ours is the first study aimed to establish a relationship between LTL and this specific stroke subtype. Furthermore, our cohort presents an additional advantage: our control group consists of patients with AF, unlike the control groups of the cited studies that were made up of healthy subjects. Thus, we have found an independent association under rather severe analytic conditions.
Also controversial is the real role played by short telomeres in cardiovascular diseases: in spite of the fact that short telomeres seem to play a key role by driving cells into senescence and apoptosis,3,4 there are studies that do not support causality.14 Our finding that the cardioembolic stroke risk displayed a dose-dependency according to LTL values would provide support for the causative hypothesis.
Our study has limitations. Its retrospective nature precludes definitive conclusions about the prognostic power of LTL in predicting the cardioembolic stroke risk. Indeed, we cannot definitely establish that telomere shortening actively plays a role in the pathogenesis of cardioembolic stroke. However, our study was performed in a relatively small population. Nevertheless, we consider that the strength and dose-dependency of the observed results provide enough arguments to establish a sound relationship between LTL and cardioembolic stroke.
In summary, our findings indicate that short LTL was independently associated with higher cardioembolic stroke risk in patients with AF. Our findings encourage prospective studies to confirm that LTL constitutes an attractive biomarker to help in the decision to start anticoagulant treatment in patients with AF.
Sources of Funding
This study was supported by grants PI08/1349, PI11/01458 and RD12/0042/0009 (Instituto de Salud Carlos III and European Regional Development Funds), Education (2008), Health (15/09), Departments of Gobierno de Navarra, Spanish Society of Thrombosis and Haemostasia (2010), Gerencia Regional de Salud de Castilla y León (GRS358/A/09), and Ministry of Science and Technology ERANET-NEURON PROTEA PRI-PIMNEU-2011-1334.
The online-only Data Supplement is available with this article at http://stroke.ahajournals.org/lookup/suppl/doi:10.1161/STROKEAHA.115.011837/-/DC1.
- Received October 13, 2015.
- Revision received November 6, 2015.
- Accepted November 17, 2015.
- © 2016 American Heart Association, Inc.
- Wolf PA,
- Abbott RD,
- Kannel WB.
- Friberg L,
- Rosenqvist M,
- Lip GY.
- Sanders JL,
- Newman AB.
- Weischer M,
- Bojesen SE,
- Cawthon RM,
- Freiberg JJ,
- Tybjærg-Hansen A,
- Nordestgaard BG.
- You NC,
- Chen BH,
- Song Y,
- Lu X,
- Chen Y,
- Manson JE,
- et al
- Ding H,
- Chen C,
- Shaffer JR,
- Liu L,
- Xu Y,
- Wang X,
- et al
- Cawthon RM.
- De Meyer T,
- Van Daele CM,
- De Buyzere ML,
- Denil S,
- De Bacquer D,
- Segers P,
- et al