Demographic determinants of REMS-derived BMD and fragility score

REMS estimates bone mineral density and fracture risk. This study found that its output is driven mainly by demographic factors (age and weight) rather than direct skeletal measurement. This affects interpretation of outliers, discordance with DXA, and monitoring. Using REMS within FRAX, which already includes demographic variables, may introduce confounding effects. Algorithm transparency is essential to reassess REMS' clinical role. PURPOSE: Recent findings have questioned the dependency of radiofrequency echographic multi-spectrometry (REMS) outputs on demographic inputs. This study investigates the extent to which REMS-derived fragility score (FS) and REMS-derived bone mineral density (REMS-BMD) values are influenced by demographic inputs (e.g., age, weight, and height) through the analysis of clinical data and controlled experimental manipulation. METHODS: A clinical cohort (178 females, 31 males) underwent REMS scans of the lumbar spine (LS) and femoral neck (FN). Multiple linear regression was used to determine the variance in REMS-BMD and FS explained by demographics. Additionally, five healthy volunteers underwent multiple REMS scans where age and weight inputs were artificially varied to measure the direct impact on the REMS-BMD output, assuming stable underlying bone. RESULTS: > 0.80) in both males and females. In the experimental cohort, artificially increasing the input age caused a FN REMS-BMD decline of approximately 6.3% per decade, while an artificial 5-kg increase in weight input caused a FN REMS-BMD increase of approximately 4.3%. CONCLUSIONS: REMS-BMD and fragility scores are strongly influenced by demographic inputs, particularly age and weight. The REMS output therefore reflects an assessment of BMD and fracture risk strongly based on demographic parameters, rather than an ultrasound-based direct skeletal measurement. The results indicate the need for caution in individual patient assessment (particularly outliers), its application with FRAX calculations, and its use for longitudinal monitoring.