Shoe–ground loading responses across three 3D-printed arch-support heights in normal- and high-arched marathon runners: an arch-type-stratified secondary analysis

Objective A prior parameter-optimization study identified 9–15-mm support heights at 22%–26% infill as a candidate range for 3D-printed arch-support inserts, but whether responses within this range differ between runners with normal and high arches remains unclear. This arch-type-stratified secondary analysis compared shoe–ground interface loading under Single and 9-, 12-, and 15-mm support conditions at 24% infill and explored the temporal localization of condition-related differences during stance. Methods The primary balanced analysis included thirty male marathon runners (15 normal arch, 15 high arch), and a sensitivity analysis included all 55 eligible runners (40 normal arch, 15 high arch). A floor-mounted Footscan® system recorded shoe–ground pressure and regional force data. Primary outcomes were maximum regional peak pressure (MaxPP), maximum metatarsal-region peak pressure (MetaPP), midfoot force impulse fraction (MIF), heel force impulse fraction (HIF), and mediolateral impulse index (MLI). Prespecified comparisons used paired or Welch t -tests with Benjamini–Hochberg false-discovery-rate correction, complemented by Arch Type × Condition linear mixed-effects models and exploratory pointwise time-series analyses. Results Under Single, none of the five primary outcomes differed between arch groups after FDR correction. In the primary balanced analysis, relative to Single, FDR-supported differences were concentrated in impulse-distribution measures: MIF increased by 17.9%–19.5% in normal-arch runners under 12–24 and 15–24 and by 18.0%–23.0% in high-arch runners across all three supports, whereas MLI decreased by 14.0%–24.0% and 12.0%–19.7%, respectively. In this analysis, no MaxPP or MetaPP comparison met the FDR threshold. In normal arches, MIF was 7.4% higher under 12–24 than 9–24 ( q = 0.041) and 9.0% higher under 15–24 than 9–24 ( q = 0.003); no direct support-condition comparison was FDR-significant in high arches. Neither the 15 between-arch change comparisons nor the five Arch Type × Condition interactions reached FDR significance in the balanced analysis, and the same 0/15 and 0/5 pattern was retained in the full-cohort sensitivity analysis. Exploratory temporal differences were localized mainly to early-to-middle stance for midfoot force share and middle-to-late stance for dynamic MLI. Conclusion In the primary balanced analysis, condition-related differences were expressed more consistently in MIF and MLI than in peak-pressure measures. Across the balanced and full-cohort analyses, there was insufficient evidence that static arch type systematically modified the height-specific or overall condition-related response pattern. The analyses did not establish overall superiority of either the 12- or 15-mm configuration. Future studies should integrate dynamic arch characteristics, comfort, and longer-term adaptation.