This study aimed to assess the accuracy of existing basal metabolic rate (BMR) prediction equations in men with chronic (>1 yr) spinal cord injury (SCI). The primary aim is to develop new SCI population-specific BMR prediction models, based on anthropometric, body composition, and/or demographic variables that are strongly associated with BMR.Methods
Thirty men with chronic SCI (paraplegic, n = 21, tetraplegic, n = 9) 35 ± 11 yr old (mean ± SD) participated in this cross-sectional study. Criterion BMR values were measured by indirect calorimetry. Body composition (dual-energy x-ray absorptiometry) and anthropometric measurements (circumferences and diameters) were also taken. Multiple linear regression analysis was performed to develop new SCI-specific BMR prediction models. Criterion BMR values were compared with values estimated from six existing and four developed prediction equations.Results
Existing equations that use information on stature, weight, and/or age significantly (P < 0.001) overpredicted measured BMR by a mean of 14%–17% (187–234 kcal·d−1). Equations that used fat-free mass (FFM) accurately predicted BMR. The development of new SCI-specific prediction models demonstrated that the addition of anthropometric variables (weight, height, and calf circumference) to FFM (model 3; r2 = 0.77), explained 8% more of the variance in BMR than FFM alone (model 1; r2 = 0.69). Using anthropometric variables, without FFM, explained less of the variance in BMR (model 4; r2 = 0.57). However, all the developed prediction models demonstrated acceptable mean absolute error ≤6%.Conclusion
BMR can be more accurately estimated when dual-energy x-ray absorptiometry–derived FFM is incorporated into prediction equations. Using anthropometric measurements provides a promising alternative to improve the prediction of BMR, beyond that achieved by existing equations in persons with SCI.