A Model of Prediction and Cross-Validation of Fat-Free Mass in Men With Motor Complete Spinal Cord Injury

Gorgey AS, Dolbow DR, Gater DR Jr. A model of prediction and cross-validation of fat-free mass in men with motor complete spinal cord injury.ObjectivesTo establish and validate prediction equations by using body weight to predict legs, trunk, and whole-body fat-free mass (FFM) in men with chronic complete spinal cord injury (SCI).DesignCross-sectional design.SettingResearch setting in a large medical center.ParticipantsIndividuals with SCI (N=63) divided into prediction (n=42) and cross-validation (n=21) groups.InterventionNot applicable.Main Outcome MeasureWhole-body FFM and regional FFM were determined by using dual-energy x-ray absorptiometry. Body weight was measured by using a wheelchair weighing scale after subtracting the weight of the chair.ResultsBody weight predicted legs FFM (legs FFM=.09×body weight+6.1; R2=.25, standard error of the estimate [SEE]=3.1kg, P<.01), trunk FFM (trunk FFM=.21×body weight+8.6; R2=.56, SEE=3.6kg, P<.0001), and whole-body FFM (whole-body FFM=.288×body weight+26.3; R2=.53, SEE=5.3kg, P<.0001). The whole-body FFMpredicted (FFM predicted from the derived equations) shared 86% of the variance in whole-body FFMmeasured (FFM measured using dual-energy x-ray absorptiometry scan) (R2=.86, SEE=1.8kg, P<.0001), 69% of trunk FFMmeasured, and 66% of legs FFMmeasured. The trunk FFMpredicted shared 69% of the variance in trunk FFMmeasured (R2=.69, SEE=2.7kg, P<.0001), and legs FFMpredicted shared 67% of the variance in legs FFMmeasured (R2=.67, SEE=2.8kg, P<.0001). Values of FFM did not differ between the prediction and validation groups.ConclusionsBody weight can be used to predict whole-body FFM and regional FFM. The predicted whole-body FFM improved the prediction of trunk FFM and legs FFM.