Author(s): BIRKLBAUER, J., PREYER, F.1, STOLLNBERGER, D.1, RUDIGIER, J.1, GONAUS, C.2, STEIDL-MÜLLER, L.3, Institution: UNIVERSITY OF SALZBURG, Country: AUSTRIA, Abstract-ID: 1652

INTRODUCTION:
Developmental factors strongly influence performance, talent identification, and training, making accurate biological age (BA) estimation essential in youth. Several time- and cost-efficient anthropometric methods exist, some without sitting height, but their validity and interchangeability remain unclear. Therefore, the present study aimed to compare commonly used anthropometric maturity-estimation methods with hand-wrist radiography as the gold-standard procedure in youth athletes and schoolchildren.
METHODS:
Anthropometric maturity-estimation methods proposed by Mirwald, Moore (with and without sitting height), Fransen, and two growth-curve-based approaches by Helsen were compared with skeletal age assessed by hand-wrist X-rays (Greulich-Pyle). The validation sample included 26 youth alpine skiers (11.7±0.7 yrs; 11 girls) and 36 children from a regular school (11.6±0.6 yrs; 17 girls). Additionally, anthropometric methods were compared among each other using 3874 datasets from U13-U14 male soccer players (12.7±0.6 yrs) and two samples of 618 boys (12.4±1.3 yrs) and 645 girls (12.3±1.3 yrs) from nine schools. As APHV- and growth-curve-based methods incorporate chronological age (CA), analyses also examined the difference between CA and BA (skeletal age).
RESULTS:
The height-weight growth-curve approach by Helsen showed the highest agreement with X-ray assessments in both sexes for BA (ICC₂,₁>.68; LoA -2.1 to 1.4 yrs) and CA-BA differences (ICC₂,₁>.57; LoA -2.2 to 1.3 yrs), despite overestimating skeletal age by >0.36 yrs (d>0.41; BF₁₀>2.1). Moore's equations, with or without sitting height, demonstrated the poorest agreement with radiography in boys (ICC₂,₁<.25; LoA -1.5 to 1.7 yrs) and girls (ICC₂,₁<.24; LoA -2.0 to 1.8 yrs). Among anthropometric methods, Fransen and Mirwald exhibited the highest agreement across all three male samples (ICC₂,₁>.95; LoA -0.4 to 0.4 yrs) and (as Fransen’s model is limited to boys) Moore and Mirwald agreed best in girls (ICC₂,₁>.86; LoA -0.4 to 0.4 yrs). Conversely, Moore and Helsen methods diverged most across all samples (ICC₂,₁<.73; LoA -1.9 to 2.4 yrs).
CONCLUSION:
Despite omitting sitting height, the Helsen method demonstrated moderate to substantial agreement with gold standard radiography, exceeding previous validation findings (e.g. Almeida-Neto et al. 2022). For youth near APHV, it offers a practical tool for determining BA with sufficient validity to enable maturity-adjusted performance evaluation. However, validity in later development remains unclear, and the method fails when height/weight exceed normative curves. The high agreements among APHV-based approaches align with Salter et al. (2021) observations and are unsurprising as methods represent inherent iterations of the original regression equation. Further validation against Tanner-Whitehouse radiography and MRI in older academy players is planned.
REFERENCES:
Almeida-Neto et al., Am J Hum Biol, 2022
Salter et al., Int J Sports Sci Coach, 2021