<b><i>Background:</i></b> Patients with premorbid functional impairment are generally excluded from acute stroke trials. We aimed to determine the impact of including such patients in the Head Positioning in acute Stroke Trial (HeadPoST) and early additional impairment on outcomes. <b><i>Methods:</i></b> Post hoc analyses of HeadPoST, an international, cluster-randomized crossover trial of lying-flat versus sitting-up head positioning in acute stroke. Associations of early additional impairment, defined as change in modified Rankin scale (mRS) scores from premorbid levels (estimated at baseline) to Day 7 (“early ΔmRS”), and poor outcome (mRS score 3–6) at Day 90 were determined with generalized linear mixed model. Heterogeneity of the trial treatment effect was tested according to premorbid mRS scores 0–1 versus 2–5. <b><i>Results:</i></b> Of 8,285 patients (38.9% female, mean age 68 ± 13 years) with complete data, there were 1,984 (23.9%) with premorbid functional impairment (mRS 2–5). A significant linear association was evident for early ∆mRS and poor outcome (per 1-point increase in ΔmRS, adjusted odds ratio 1.20, 95% confidence interval 1.14–1.27; <i>p</i> &#x3c; 0.0001). Patients with greater premorbid functional impairment were less likely to develop additional impairment, but their risk of poor 90-day outcome significantly increased with increasing (worse) premorbid mRS scores (linear trend <i>p</i> &#x3c; 0.0001). There was no heterogeneity of the trial treatment effect by level of premorbid function. <b><i>Conclusions:</i></b> Early poststroke functional impairment that exceeded premorbid levels was associated with worse 90-day outcome, and this association increased with greater premorbid functional impairment. Yet, including premorbid impaired patients in the HeadPoST did not materially affect the subsequent treatment effect. <b><i>Clinical Trial Registration:</i></b> HeadPoST is registered at http://www.ClinicalTrials.gov (NCT02162017).