Comparative outcomes following primary first metatarsophalangeal arthrodesis using CrossCheck, IOFIX and ORTHOLOC3DI constructs: a multicentre retrospective cohort study

Background First metatarsophalangeal joint (MTPJ) arthrodesis is the gold-standard treatment for end-stage hallux rigidus, inflammatory arthropathy and selected hallux valgus. This study compared radiographic union following first MTPJ arthrodesis using three fixation constructs: the CrossCheck dorsal locking plate, the IOFIX intraosseous fixation system and the ORTHOLOC3DI dorsal locking plate. Methods This multicentre retrospective cohort study included primary first MTPJ arthrodeses performed between October 2021 and March 2026 at three orthopaedic units in Manchester, UK; implant selection was at the operating surgeon’s discretion. The primary outcome was radiographic union; secondary outcomes were time to union and postoperative complications. Union rates were compared using Fisher’s exact test with exact 95% confidence intervals (CI), and time to union using the Kruskal–Wallis test with Bonferroni-adjusted pairwise Mann–Whitney U tests. Results Sixty-four procedures in 63 patients were included, and union was achieved in 61 (95.3%). Union rates were 100% (28/28; 95% CI 87.7–100) for CrossCheck, 94.1% (16/17; 95% CI 71.3–99.9) for ORTHOLOC3DI and 89.5% (17/19; 95% CI 66.9–98.7) for IOFIX ( p = 0.171). Complications occurred after 11 procedures (17.2%), with no significant differences between constructs. Degenerative pathology accounted for 89.3% of CrossCheck procedures, compared with 31.6% of IOFIX and 58.8% of ORTHOLOC3DI procedures ( p < 0.001). Median time to union was shorter with CrossCheck than with ORTHOLOC3DI (43 [IQR 43–50] vs. 51 [IQR 45–83.5] days; Bonferroni-adjusted p = 0.038), but this difference was not statistically significant when the analysis was restricted to degenerative pathology ( p = 0.217), although this subgroup analysis had limited statistical power. Conclusion No statistically significant difference in union rates was identified between the three constructs; given the small number of non-union events, these data cannot establish equivalence or non-inferiority. The shorter time to union with CrossCheck was not maintained in degenerative pathology, and implant choice was strongly confounded by surgical indication and surgeon preference. These preliminary, hypothesis-generating findings require confirmation in adequately powered, indication-matched prospective studies.