NIHSS cutoff scores for predicting FIM motor severity at discharge in acute stroke

Early estimation of activities of daily living outcomes after acute stroke is important for rehabilitation and discharge planning. The aims were to (1) examine Functional Independence Measure (FIM) motor, cognitive, and total score distributions and floor and ceiling effects at discharge; (2) determine the incremental explanatory value of the acute-phase National Institutes of Health Stroke Scale (NIHSS); and (3) derive and internally validate NIHSS cutoff rules for distinguishing adjacent FIM motor severity levels and evaluate their incremental classification value. This retrospective cohort included 974 patients; cognitive and total FIM analyses included 969. Regression analyses and a stratified 70:30 derivation-validation split were used. Ceiling effects were observed for the FIM motor (19.2%), cognitive (34.7%), and total scores (16.0%), whereas no floor effects were observed. Adding NIHSS to clinical variables increased the explained variance for the FIM motor (ΔR² = 0.253), cognitive (ΔR² = 0.332), and total scores (ΔR² = 0.292; all p < .001). Higher NIHSS scores, older age, higher pre-stroke modified Rankin Scale scores, and hemorrhagic stroke were associated with greater FIM motor severity. The NIHSS >17 and >5 rules distinguished severe from moderate and moderate from mild severity, with validation-sample AUCs of 0.777 and 0.780. Adding these rules to baseline models improved the validation-sample AUCs for both comparisons, although paired classification accuracy improved significantly only for severe versus moderate severity. Sensitivity was below 0.70 in some analyses. NIHSS cutoff rules may provide reference points for estimating FIM motor severity at discharge but should not be used as stand-alone predictors.