Telomere length and stroke recurrence after ischemic stroke and TIA

Abstract

Background and Objective: Shortening telomere length (TL), as an indicator of aging, has been associated with increased risk of cardiovascular disease and incident stroke. However, there are limited data relating to the association between TL and recurrent stroke. Methods: Patients from the Third China National Stroke Registry who had whole genome sequencing (WGS) were selected. TL was estimated using TelSeq based on binary sequence alignment/map files derived from WGS data. Cox proportional hazards regression models were performed to assess the association of TL with recurrent stroke. Results: A total of 8041 patients with ischemic stroke (IS) or transient ischemic attack (TIA) were included. Mean TL was 2.14 ± 0.82 kb. Patients in the lowest tertile of TL had higher incidence of stroke recurrence compared to those in the middle and highest tertile (6.4% vs 5.9% vs 5.2%), but the difference was not longer significant after adjusting for age, sex, cardiovascular risk factors and stroke severity. Similarly, when analyzing TL as a continuous variable, the HR per 1000 bp increase in TL was significant 0.88 (0.79–0.98), but after adjusting for co-variates, was no longer significant (0.91; 95% confidence interval (CI), 0.81–1.02). In patients aged > 65 years, but not in younger patients, after adjusting for co-variates, TL was significantly associated with stroke recurrence. Compared to the lowest tertile, HRs (95% CI) after adjustment for all co-variates for the middle and highest tertiles were 0.78 (0.55–1.10) and 0.67 (0.46–0.98), respectively, with p for trend of 0.03. In analyses using TL as a continuous variable, adjusted HR (95% CI) per 1000 bp increase in TL was 0.80 (0.66–0.96). However, there was no significant interaction between TL and age on risk of stroke recurrence (p for interaction = 0.09). Conclusions: In Chinese IS or TIA patients, no independent association was found between TL and risk of stroke recurrence after adjusting for co-variates. We found a possible association in older patients but this needs replicating.