Logo image
Back
Time-Varying Predictors of Stroke Following Transient Ischaemic Attack: A National Population-Based Cohort Study Using 50 Years of Linked Healthcare Data in Wales, UK
Preprint

Time-Varying Predictors of Stroke Following Transient Ischaemic Attack: A National Population-Based Cohort Study Using 50 Years of Linked Healthcare Data in Wales, UK

Renin Toms, Jessica O Williams, Hywel Evans, Alexander Deardon, Keith Morris and Philip E James
Preprints with The Lancet, Vol.12 June 2026
Elsevier
2026
DOI: https://doi.org/10.2139/ssrn.6911553
url
https://dx.doi.org/10.2139/ssrn.6911553View
Preprint Version Open

Abstract

Transient ischaemic attack Stroke risk Time-varying analysis Secondary prevention Routinely collected health data Population-based cohort Coagulation testing Wales
Background: Transient ischaemic attack (TIA) confers a high risk of subsequent stroke. Population-level evidence describing how stroke risk after TIA evolves over time is needed to inform effective secondary prevention. Methods: We conducted a national population-based retrospective cohort study of adults with a recorded TIA event in Wales. Time to first subsequent stroke was modelled using multivariable Cox proportional hazards regression and follow-up was split into prespecified intervals (0–7, 8–30, 31–365, and >365 days) to estimate interval-specific associations. Sensitivity analyses were conducted using a generalised linear mixed-effects model (GLMER). Findings: Among 40,957 individuals with a recorded TIA event, 10,572 (25.8%) had a subsequent stroke. Stroke risk since TIA varied over time. Increasing age was associated with progressively higher stroke risk, particularly beyond one year (75–84 years vs 18–44 years: HR 2.25, 95% CI 1.80–2.81). Diabetes was associated with persistently elevated stroke risk across all follow-up periods (0–7 days: HR 1.29, 95% CI 1.14–1.46; >365 days: HR 1.27, 95% CI 1.15–1.41). Elevated cardiovascular risk (QRISK2 ≥10) was associated with increased stroke risk in the early post-TIA period (HR 2.20, 95% CI 1.30–3.72), while unavailable QRISK2 scores were associated with higher risk, beyond one year (HR 4.04, 95% CI 2.42–6.74). Socioeconomic deprivation was associated with gradient increase in longer-term stroke risk especially among the most deprived. Standard coagulation parameters were recorded in only 16.2% of patients. In sensitivity analyses, individuals without any recorded coagulation-related testing had higher odds of subsequent stroke (aOR 1.48, 95% CI 1.38–1.59). Interpretation: Stroke risk following TIA is dynamic, with key predictors showing important time-varying associations. These findings support time-aware risk stratification, enhanced post-TIA monitoring, and the need for and sustained post-TIA care in longer periods.

Details

Metrics

1 Record Views
Logo image