Ageing is associated with weight loss and subsequently poor health outcomes. The present study assessed agreement between two field methods, bioelectrical impedance spectroscopy (BIS) and corrected arm muscle area (CAMA) for assessment of body composition against dual-energy X-ray absorptiometry (DXA), the reference technique. Agreement between two predictive equations estimating skeletal muscle mass (SMM) from BIS against SMM from DXA was also determined. Assessments occurred at baseline < 14 d post-surgery (n 79), and at 6 months (6M; n 75) and 12 months (12M; n 63) in community-living older adults after surgical treatment for hip fracture. The 95 % limits of agreement (LOA) between BIS and DXA, CAMA and DXA and the equations and DXA were assessed using Bland-Altman analyses. Mean bias and LOA for fat-free mass (FFM) between BIS and DXA were: baseline, 0.7 (-10.9, 12.4) kg; 6M, - 0.5 (-20.7, 19.8) kg; 12M, 0.1 (-8.7, 8.9) kg and for SMM between CAMA and DXA were: baseline, 0.3 (-11.7, 12.3) kg; 6M, 1.3 (-4.5, 7.1) kg; 12M, 0.9 (-5.4, 7.2) kg. Equivalent data for predictive equations against DXA were: equation 1: baseline, 15.1 (-9.5, 20.6) kg; 6M, 17.1 (-12.0, 22.2) kg; 12M, 17.5 (-13.0, 22.0) kg; equation 2: baseline, 12.6 (-7.3, 19.9) kg; 6M, 14.4 (-9.7, 19.1) kg; 12M, 14.8 (-10.7, 18.9) kg. Proportional bias (BIS: β = -0.337, P< 0.001; CAMA: β = -0.294, P< 0.001) was present at baseline but not at 6M or 12 M. Clinicians should be cautious in using these field methods to predict FFM and SMM, particularly in the acute care setting. New predictive equations would be beneficial.