Ken Nishimura, Cameron Ward, Mehmet Furkan Tunaboylu, Andrew F Emanuels, Steven L Moran
In this small, selected cohort, rater-averaged radiographic WI was numerically lower in wrists with nonunion, suggesting a preliminary, hypothesis-generating WI-nonunion signal. The conservative scaphoid type I versus type II dichotomization was imprecise and should not be interpreted as definitive. Larger studies with more complete follow-up and adjustment for technical and clinical confounders are needed.
PURPOSE: To evaluate whether lower radiographic scaphoid waist index (WI), measured on routine preoperative lateral wrist radiographs, is associated with nonunion after scaphocapitate fusion for Kienböck disease.
METHODS: We performed a retrospective chart and imaging review of patients who underwent scaphocapitate fusion for Kienböck disease within a single integrated multicampus health system (1996-2025). The primary outcome was nonunion. Symptomatic adjacent-joint osteoarthritis within 1 year was evaluated separately as a descriptive secondary outcome. WI was measured independently by two observers on preoperative lateral radiographs. Rater-averaged WI was analyzed as a continuous variable. A conservative dichotomized scaphoid type analysis was performed exploratorily, with type II assigned only when both observers measured WI < 0.40.
RESULTS: Of 98 scaphocapitate fusions identified, 22 met inclusion criteria after excluding cases with nonretrievable archived imaging or follow-up <1 year. Median available postoperative follow-up was 23.7 months (interquartile range [IQR]: 14.4-41.2 months); range, 12.0-185.4 months. Nonunion occurred in 5 of 22 wrists (23%). Rater-averaged WI was numerically lower in wrists with nonunion than in wrists without nonunion: median 0.365 (IQR: 0.364-0.365) versus 0.429 (IQR: 0.375-0.452). The Hodges-Lehmann median difference was -0.0568, with a bootstrap 95% CI: -0.0981 to 0.0087. In the exploratory conservative dichotomized analysis, nonunion occurred in four of nine scaphoid type II wrists and 1 of 13 scaphoid type I wrists; the odds ratio estimate was imprecise and included the null (OR: 9.60; exact 95% CI: 0.64-510.98). Symptomatic adjacent-joint osteoarthritis occurred in two wrists.
CONCLUSIONS: In this small, selected cohort, rater-averaged radiographic WI was numerically lower in wrists with nonunion, suggesting a preliminary, hypothesis-generating WI-nonunion signal. The conservative scaphoid type I versus type II dichotomization was imprecise and should not be interpreted as definitive. Larger studies with more complete follow-up and adjustment for technical and clinical confounders are needed.
TYPE OF STUDY/LEVEL OF EVIDENCE: Prognostic III.