This study showed that BIVA outcomes including PhA, height-adjusted R (R/H), and Xc (Xc/H), were related to BMD levels in the whole body, spine, and proximal femur in the elderly population. This provided a new biomarker for BMD using BIVA. To our knowledge, only one study has related PhA as a BMD determinant and no study has related vector shift in the RXc graph with the changes in BMD. In the study by Antunes et al. (18), a significant correlation (p<0.001) exists between PhA and BMD in the total body, femur, femoral neck, and forearm with correlation coefficients of 0.475, 0.524, 0.450, and 0.437, respectively, but no correlation exist in the spine. The cause of the discrepancy remains unclear. This study is consistent with Antunes et al. showing a moderately positive correlation between BMD and PhA.
The phase angle is the most established BIVA parameter. It is considered to be age- and gender-dependent27. When adding body composition parameters as variables, age explained most of the variability in PhA, followed by the fat-free mass (FFM) in healthy subjects28. In a population study, PhA was found to be associated with frailty and mortality, independent of age and comorbidity17,29. Additionally, FFM is positively correlated with BMD with varying strength of the correlation, depending on the sites of BMD30,31. Therefore, it is reasonable to expect that PhA may be correlated with BMD. However, more research is needed to explore whether potential confounding variables such as age, gender, height, and weight can affect the correlational relationship between PhA and BMD in different body regions.
R and Xc are BIVA parameters directly measured from the BIA device. By normalized to body height, R/H and Xc/H can provide a qualitative measure of soft tissue that does not depend on body size20. This study showed that BMD was positively correlated with Xc/H but negatively correlated with R/H. Each unit increased in Xc/H (Ohm/m) was associated with 0.0066 g/cm2 increase in BMDtotal whereas each unit increased in R/H (Ohm/m) was associated with a 0.0016 g/cm2 decrease in BMDtotal. A similar trend was observed for estimating BMD in the spine and proximal femur in this study. This is the first study suggesting that BIVA variables might be predictors for BMD in the elderly population. Interestingly, this trend has also been observed in previous studies where muscle strength was positively correlated with Xc/H and negatively correlated with R/H8,10. Muscle strength is a positive predictor for BMD 32,33, and this might explain why R/H and Xc/H predict both muscle strength and BMD in a similar direction.
In the BIVA plot, a vector has angle as well as magnitude information in the R/H and Xc/H directions. In contrast, a PhA only has the directional component without providing information on its distance. Indeed, vectors of varying length share the same PhA in the RXc graph. Therefore, the vector analysis approach has the potential of presenting more diverse information on biological activity compared to PhA alone. In addition, the BIA device is a reliable tool, providing measurement with a precision error of 2%-3% 34,35. Currently, BIVA has been used to differentiate obesity (higher PhA and short length), athletes (higher PhA, long length), cachexia (lower PhA, long length), and weakens (normal PhA and short length). In this study, higher BMD was associated with a vector of higher PhA and shorter length. This may provide a diagnostic tool for estimating BMD in the elderly population at risk of osteoporosis.
The present study showed that PhA, R/H, and Xc/H estimates were correlated with BMD. A low PhA is associated with osteoporosis, despite age and gender control18,36. BIA device is well-known for personal body fat monitoring at home. However, the current home-use models do not provide BIVA estimates. With the development of home-use BIA devices providing BIVA estimates, a self-screening tool for osteoporosis may be created, aiding in the early diagnosis and treatment of osteoporosis.
This study has several limitations. First, a cause-effect relationship between BMD and height-adjusted R and Xc could not be established. Second, subgroup analysis was not performed due to the small sample size. Third, osteoporosis is a complex disease governed by a wide range of phenotypes regulated by genetic and environmental factors37. However, these factors such as age, gender, ethnicity, family history, physical activity were not included as variables in the regression analysis. Further study is still needed to investigate the impact of these factors on BMD.