Handgrip strength beyond binary thresholds: classification structure and graded functional associations in stroke and hip fracture.

BACKGROUND

International handgrip strength (HGS) criteria use different thresholds and may classify the same patient differently. We examined classification structure across six criteria and the graded association of admission HGS with functional recovery after stroke and hip fracture.

METHODS

We retrospectively analyzed patients aged ≥65 years admitted to convalescent rehabilitation wards after stroke or proximal femoral fracture. Six sex specific AWGS 2025, EWGSOP2, FNIH, and SDOC criteria were applied. The primary outcome was eligible-item independence recovery, defined as a Functional Independence Measure (FIM) motor item scored 1-5 at admission reaching 6-7 at discharge. Disease-specific logistic generalized estimating equations used inverse cluster-size weighting. Patient-level FIM motor gain was examined as a clinical benchmark.

RESULTS

Threshold-comparison cohorts included 929 patients with stroke and 774 with hip fracture. Low-strength prevalence ranged from 39.0% to 87.0% after stroke and from 34.2% to 92.9% after hip fracture. The four absolute criteria yielded nested classifications as expected from their ordered cut-points, whereas BMI-normalized criteria produced the principal departures from this structure. Each 5-kg higher admission HGS was associated with 3.3 (95% CI, 1.9-4.7) and 6.4 (4.4-8.5) additional recoveries per 100 eligible items after stroke and hip fracture, respectively. Patient level FIM motor gain showed concordant associations. Recorded 0-kg HGS occurred in 15.4% and 1.8% of age-eligible patients, respectively, and was treated as a separate measurement state.

CONCLUSIONS

International HGS thresholds are criterion-specific classification anchors rather than interchangeable biological boundaries. Exact HGS values retain graded functional information beyond binary classification, while testability and measurement state should be documented separately.