Critical Care Decisions After Large Core Cerebral Infarctions: A Secondary Analysis From the SELECT2 Trial

Among patients with large vessel occlusion (LVO) and large ischemic cores, critical decisions often need to be made about decompressive hemicraniectomy (DHC) or early withdrawal of life-sustaining therapy (WLST). In this study, we aimed to evaluate utilization of DHC and early WLST and factors associated with them in patients with large strokes from the SELECT2 trial.

We analyzed the entire SELECT2 trial population, which randomized 352 patients with stroke due to LVO and large ischemic cores to endovascular thrombectomy (EVT) or medical management. We used the as-treated principle to compare the use of DHC and early WLST within 7 days after randomization. We further assessed functional outcomes (modified Rankin Score) after these decisions.

Of 352 patients enrolled in this study, 55 received DHC and 81 transitioned to early WLST. Patients treated with EVT were as likely to undergo DHC (16% vs 15%, adjusted relative risk [aRR] = 1.19, 95% CI:0.75-1.88, p = 0.46) or WLST (22% vs 24%, aRR = 0.94, 95% CI: 0.66-1.34, p = 0.72) as those given medical management. DHC was used more frequently in younger patients and WLST more in older patients. EVT efficacy was maintained after adjusting for DHC (adjusted generalized odds ratio [aGenOR] = 1.68, 95% CI: 1.24-2.11, p < 0.001), with no interaction between DHC and treatment (p-interaction = 0.93). At 1 year, 21% of DHC-treated patients were ambulatory; the outcomes were universally poor after early WLST.

In the SELECT2 trial of patients with large ischemic core, DHC was performed in ~1 of 6 patients and early WLST in ~1 of 5 patients, without differences based on treatment with EVT or medical management, nor successful reperfusion. DHC or WLST did not detract from thrombectomy treatment benefit. Additionally, ~20% of patients achieved independent ambulation despite receiving DHC by the 1-year follow-up. The similar distribution of these critical care decisions provides reassurance that the overall trial outcomes were not biased by open-label treatment allocation. ANN NEUROL 2025;97:698-708.