Apraxia as a Predictor of Poststroke Recovery: Insights From the Birmingham Cognitive Screening Program

Abstract

Background: Limb apraxia is common after stroke and may affect long-term activities of daily living. This study investigates whether early subacute limb praxis scores predict long-term activities of daily living outcomes over and above other cognitive deficits.

Methods: This longitudinal observational study analyzed data from the BCoS (Birmingham Cognitive Screen) cohort, conducted between 2010 and 2015 across multiple stroke centers in the United Kingdom. Two-hundred-fifty-six first-ever computed tomography confirmed stroke survivors (56.3% men; mean±SD age=68.3±11.4 y) were assessed <1 month poststroke (subacute) and >9 months (chronic). BCoS cohort comprises 34 cognitive tasks and 4 assess limb praxis. Scores were rescaled to 0 to 100. Functional outcome was the 20-point Barthel index (BI). Stepwise multiple linear regression with 4-fold internal cross-validation tested whether subacute cognitive and praxis performance predicted chronic BI, adjusting for baseline BI, and other cognitive domains.

Results: Mean BI improved from 13.3±5.5 to 17.3±3.9. Higher subacute limb praxis scores predicted better chronic activities of daily living: gesture production β=-0.0555 (P=0.0008), gesture recognition β=-0.0349 (P=0.017), meaningless gesture imitation β=0.0338 (P=0.0047). The full model explained 60% of BI variance and outperformed a model without praxis measures (ΔR²=0.04; ANOVA P<0.001).

Conclusions: Detailed early limb praxis testing adds independent prognostic value for long-term activities of daily living and should be incorporated into routine poststroke assessments to target rehabilitation.

Keywords: activities of daily living; apraxias; gesture; morbidity; stroke.

Figure.

Associations between early subacute praxis performance and change in activities of daily living (ADL). Each shows a linear fit (solid line) with 95% CI (shaded) for the relation between Barthel index chronic (ΔBI) ADL–subacute, range (0–20) and (A) meaningless gesture imitation (%), (B) gesture production (%), and (C) gesture recognition (%). Points are jittered for visibility. Praxis scores are scaled 0 to 100.