BackgroundIdentifying muscle weakness is challenging, because the reduction of strength with aging does not depend only on sarcopenia, but also on sensorimotor deficits. Nevertheless, this identification is improved by adjusting muscle strength allometrically, by removing the influence of body size. However, the effectiveness of foreign models to normalize these (dys)functionalities is not yet tested. This study aimed to compare and apply foreign allometric exponents for normalizing isokinetic knee extension strength in Portuguese older adults to identify muscle weakness/mobility limitation. Additionally, to attest any populational difference, data of these people and Brazilian older adults were comparedMethodsThis is a cross-sectional study encompassing 226 Portuguese (n = 132) and Brazilian (n = 94) older adults. Mobility limitation (six-minute walk test, at lowest quartile), lower limb strength (knee extension isokinetic strength at 60º/s), and body dimensions measures were taken. Foreign allometric exponents (b) were used to normalize Portuguese strength (strength/body-size variablesb). Non-normalized and normalized strength were compared (ROC) to generate the most accurate cut-point for identifying muscle weakness/mobility limitation.ResultsOlder Portuguese men and women had better mobility than their Brazilian counterparts. Older Portuguese women had superior muscle strength to Brazilian women. Normalization from 11 foreign models removed the influence of body size on muscle strength, with a negligible correlation (r ≤ 0.30). In contrast to the non-normalized strength, the normalized strength cut-off points were sufficiently accurate (AUC ≥ 0.70) to avoid identifying false-negative cases of weakness/mobility limitation.ConclusionsPortuguese older women were stronger and had superior functional capacity compared to Brazilian ones. Normalized foreign models improved the accuracy in identifying muscle weakness/mobility limitation in Portuguese older adults. The isokinetic knee extension muscle strength normalized, even using foreign allometric exponents, should be better than no adjustment.