External validation of population specific formulae for estimation of fat mass compared with DEXA measurements in children with inflammatory bowel disease

Introduction: Inflammatory bowel disease (IBD) patients have altered body composition compared to healthy controls. Dual energy X-ray absorptiometry (DEXA), the gold standard for body composition analysis, is not always available in clinical practice. In 2021 we developed population specific predictive formulae (PF) based on anthropometry for estimation of fat mass (FM%) in IBD children. We aimed to validate PF compared with DEXA. Methods: From 09/2023 to 03/2024 we prospectively enrolled IBD children at “V. Buzzi” Hospital. All patients underwent anthropometric evaluation and DEXA scan. We performed Bland-Altman analysis and calculation of percentage error (PE) between values of FM% estimated with formulae and measured by DEXA Results: 29 IBD patients were enrolled (21 Crohn’s Disease, 8 ulcerative colitis). Mean PE was 10.4% (95%CI 6.4/15.5) for PF based on BMI, 12.1% (95%CI 5.4/18.4) for PF based on triceps skinfold, 12.3% (95%CI 6.4/14.4) for PF based on two skinfolds (biceps+triceps), and 5.9% (95%CI 4.4/9.6) for PF based on 4 skinfolds (SF4, triceps+biceps+subscapular+suprailiac). The smallest range between the limits of agreement of Bland-Altman analysis was found for SF4: lower limit -6.2 (95%CI -8.2/-4.3) and upper limit 4.6 (95%CI 2.6/6.5). Conclusion: The most accurate PF for estimation of FM% in paediatric IBD patients is SF4 (triceps + biceps + subscapular+ suprailiac) which shows the lowest PE and the best agreement between methods. This formula may be a valid tool for estimation and monitoring of nutritional status and a valid alternative to DEXA measurement.