Association between age and risk of stroke or death from carotid endarterectomy and carotid stenting: a meta-analysis of pooled patient data from four randomised trials

Background: Age was reported to be an effect-modifier in four randomised controlled trials comparing carotid artery stenting (CAS) and carotid endarterectomy (CEA), with better CEA outcomes than CAS outcomes noted in the more elderly patients. We aimed to describe the association of age with treatment differences in symptomatic patients and provide age-specific estimates of the risk of stroke and death within narrow (5 year) age groups.

Methods: In this meta-analysis, we analysed individual patient-level data from four randomised controlled trials within the Carotid Stenosis Trialists' Collaboration (CSTC) involving patients with symptomatic carotid stenosis. We included only trials that randomly assigned patients to CAS or CEA and only patients with symptomatic stenosis. We assessed rates of stroke or death in 5-year age groups in the periprocedural period (between randomisation and 120 days) and ipsilateral stroke during long-term follow-up for patients assigned to CAS or CEA. We also assessed differences between CAS and CEA. All analyses were done on an intention-to-treat basis.

Findings: Collectively, 4754 patients were randomly assigned to either CEA or CAS treatment in the four studies. 433 events occurred over a median follow-up of 2·7 years. For patients assigned to CAS, the periprocedural hazard ratio (HR) for stroke and death in patients aged 65-69 years compared with patients younger than 60 years was 2·16 (95% CI 1·13-4·13), with HRs of roughly 4·0 for patients aged 70 years or older. We noted no evidence of an increased periprocedural risk by age group in the CEA group (p=0·34). These changes underpinned a CAS-versus CEA periprocedural HR of 1·61 (95% CI 0·90-2·88) for patients aged 65-69 years and an HR of 2·09 (1·32-3·32) for patients aged 70-74 years. Age was not associated with the postprocedural stroke risk either within treatment group (p≥0·09 for CAS and 0·83 for CEA), or between treatment groups (p=0·84).

Interpretation: In these RCTs, CEA was clearly superior to CAS in patients aged 70-74 years and older. The difference in older patients was almost wholly attributable to increasing periprocedural stroke risk in patients treated with CAS. Age had little effect on CEA periprocedural risk or on postprocedural risk after either procedure.

Funding: None.