Impact of abdominal adiposity correction on trabecular bone score (TBS) in obese women: A comparative study of software versions 3.0 and 4.0 with a predictive model.

Abstract

To evaluate the impact of soft-tissue thickness correction on trabecular bone score (TBS) in obese women across different BMI categories, comparing software versions 3.0 and 4.0, and to examine their associations with bone mineral density (BMD), fracture occurrence, and clinical variables. This cross-sectional study included 247 obese women (BMI ≥30 kg/m²) who underwent lumbar spine densitometry with TBS assessment using versions 3.0 and 4.0 of the TBS iNsight® software. Demographic, anthropometric, and BMD data were analyzed. Correlations and multivariate linear regression models were used to assess associations between TBS 4.0, BMD, fractures, and clinical parameters. TBS 4.0 were lower than TBS 3.0 across all BMI categories (p < 0.001), with the largest reductions observed in women with BMI ≥ 37 kg/m². Classification of TBS changed in 34% of cases when using version 4.0, with a greater proportion reclassified to degraded categories. TBS 4.0 showed a positive correlation with lumbar spine BMD (r = 0.51, p < 0.001) and BMI (r = 0.17, p = 0.007), but showed no significant correlation with abdominal wall thickness (r = -0.08, p = 0.229). The correlation between TBS and BMI was stronger for version 3.0 than for version 4.0 (r = 0.32 vs. r = 0.17, respectively). In multivariate analysis, abdominal wall thickness emerged as an independent predictor of TBS 4.0. A multivariate regression model including TBS 3.0, age, L1-L4 BMD, and abdominal wall thickness explained 86% of the variability in TBS 4.0, enabling estimation of TBS 4.0 from clinical variables and the earlier software version. Women with fractures showed lower TBS values, though not significant. The TBS 4.0, by incorporating automatic correction for soft-tissue thickness, provides lower values that possibly better characterize trabecular microarchitecture in individuals with obesity, especially in those with higher BMIs, and mitigates the bias attributable to differences in body composition observed with previous versions of the TBS software.