Abstract / Summary
Evidence guiding antithrombotic escalation after early neurological deterioration (END) remains limited. This single-center retrospective cohort study evaluated early neurological, exploratory 90-day functional, and safety outcomes following post-END rescue antithrombotic strategies in acute ischemic stroke. Consecutive adults admitted between January 2020 and December 2021 with acute ischemic stroke and END were classified by their final post-END strategy: continued dual antiplatelet therapy (DAPT), tirofiban, argatroban, or sequential tirofiban-to-argatroban rescue escalation. Multivariable logistic regression assessed 7-day neurological improvement, defined as a decrease of at least 4 points on the National Institutes of Health Stroke Scale (NIHSS) from END recognition; 90-day functional outcomes were exploratory. Sequential-rescue estimates were retained for descriptive completeness because delayed escalation introduces immortal-time and survivor-selection bias. Improvement rates were also summarized descriptively by TOAST subtype. Among 259 patients, 111 received continued DAPT, 54 tirofiban, 65 argatroban, and 29 sequential rescue. Seven-day neurological improvement occurred in 36.9%, 61.1%, 49.2%, and 69.0%, respectively (overall P = 0.003). In the four-category adjusted model, the OR was 2.74 (95% CI 1.36–5.51; P = 0.005) for tirofiban and 1.84 (95% CI 0.94–3.59; P = 0.075) for argatroban versus continued DAPT. Sequential-rescue findings were descriptive only. Substantial TOAST imbalance persisted after overlap weighting (maximum pairwise |SMD| = 0.654). Documented drug-monitoring/laboratory abnormalities occurred in 0.0%, 1.9%, 0.0%, and 13.8%, respectively (exact P < 0.001). Within the final-strategy analysis, tirofiban was associated with higher odds of 7-day neurological improvement than continued DAPT, whereas the argatroban estimate remained uncertain. Favorable outcomes observed after sequential rescue are descriptive only and do not establish a treatment effect. Residual confounding, including substantial TOAST imbalance, precludes claims of treatment superiority; nonstandardized surveillance also limits safety comparisons. Not applicable.