Muscle strength and ageing: methodological aspects of isokinetic dynamometry and androgen administration

1. To evaluate interventions aiming to increase muscle strength in older men, it is necessary to use an objective and reproducible method to measure strength. The most reliable method to evaluate muscle strength is isokinetic peak torque (PT) measured by a dynamometer; however, raw PT varies with differences in body size and an optimal scaling for body physique to control confounding effects of body size is necessary to make the most valid comparisons. 2. The present study was designed to estimate the effects of age (part A) and androgen administration (part B) on muscle strength and to estimate reproducibility and evaluate various scaling methods in order to optimize comparisons of isokinetic PT measurements. 3. A single isokinetic exercise protocol was used to compare the muscle strength of 31 healthy men of two age groups (< 40 and > 60 years; part A) and change in strength due to administration of dihydrotestosterone (DHT; 70 mg/day) or placebo gel for 3 months in 35 healthy older (> or = 60 years) men (part B). 4. Muscle strength was assessed by a total of 16 PT measurements using a Cybex NORM dynamometer (Cybex, Ronkonkoma, NY, USA). Age-related differences in muscle strength were estimated by using PT evaluated as raw data or scaled by normalizing methods, including simple ratio (PT/weight), allometric PT (PT/weight(0.67)) and adjustment of PT by weight, height, body mass index and body surface area (BSA). The goodness-of-fit for various scaling methods was compared using the Akaike Information Criterion (AIC) as an objective measure of model-based entropy reduction. 5. The effects of DHT administration according to different scaling methods were estimated by eta-squared measure of effect size in treatment models. In part A, older men were weaker than younger men in five knee PT, consistently by all eight analysis models but not in shoulder PT. In part B, DHT treatment resulted in an increase one knee PT (dominant knee flexion at 120 degrees/s) with the difference consistent in all seven models. 6. The scaling model using BSA proved superior to other comparison models throughout both parts of the present study according to entropy minimization criteria (AIC) for goodness-of-fit of the model or eta-squares for treatment effect size. 7. We conclude that differences in muscle strength due to age or androgen administration in older men are restricted to a minority of lower limb contractions and that use of BSA scaling for PT values is considered the best scaling method for muscle strength comparisons in either cross-sectional or longitudinal studies.