Success of Scaphoid Nonunion is Independent of Proximal Pole Vascularity

Abstract

Grant received from: Orthopaedic Research and Education Foundation Resident Research Project Grant #15-038 Royalty: Elsevier; Extremity Medical (Wolfe) Consulting Fee: Coventus, Trimed (Wolfe) Scaphoid vascularity following nonunion is of considerable concern when planning screw fixation and is thought to correlate with likelihood of healing. We hypothesized that greater proximal pole vascularity would correspond with increased union rate and shorter time to union. Thirty-six scaphoid patients in a prospective registry were followed longitudinally. All nonunions were treated with curettage, nonvascularized autogenous grafting, and headless screw fixation. Preop MRIs in 23 were evaluated for proximal pole vascularity by a musculoskeletal radiologist according to a 3-point system. Bleeding points were assessed intraoperatively according to a 3-point scale. Proximal pole fragments were graded for presence of necrotic trabeculae (yes/no) and remodeling potential as evidenced by osteoblastic activity (none, focal, robust) by blind assessment of 2 independent musculoskeletal pathologists. Healing was assessed on CT scan as > 50% bony bridging. Chi-square and non-parametric Kruskal Wallis tests were used to evaluate associations between diagnostic tests. Non-parametric Mann Whitney U and Spearman’s rho tests were used to compare scoring for each diagnostic modality with healing. Agreement between pathologists was assessed with weighed Kappa analysis. Thirty-four (94%) patients healed by average 13 weeks. One experienced delayed union secondary to hardware failure and healed 18 weeks following revision ORIF. Another patient is at 22 weeks with gradual bony bridging. Thirty-nine percent of proximal poles demonstrated ischemia by MRI criteria, but none were infarcted. Fifteen percent had poor punctate bleeding intraoperatively. Forty-five percent proximal poles demonstrated > 50% necrosis on histopathologic analysis yet 91% demonstrated forcal or robust remodeling. Inter-observer agreement for pathology viability was 0.667. There was no significant association between time to union and histologic viability (P = 0.349), remodeling potential (P = 0.704), bleeding points (P = 0.959), MRI global viability (P = 0.507), or MRI signal (P = 0.419). There was no association between mean MRI viability or signal and pathology grading (P = 0.224, P = 0.743). There was also no association between MRI viability or signal and intraoperative bleeding (P = 0.364, P = 0.386), or between pathologic grading and intraoperative bleeding (P = 0.904). (Figs. 59-1, 59-2) •There was no correlation between time to union and any diagnostic modality. MRI and histology did not demonstrate a correlation, indicating that these two assessments measure different aspects of the dysvascular response to injury.•Despite pathological evidence of dysvascular bone in nearly half of our patients, removal of necrotic bone and non-vascularized bone grafting with rigid internal fixation led to healing in the overwhelming majority of cases.•We conclude that true proximal pole infarction is a decidedly rare occurrence, and that vascularized bone grafting is seldom required.Figure 59-2T2-weighted coronal proton-density MR image of the scaphoid proximal pole demonstrates foci of devitalized bone marrow and ischemia, but retained vascularity.View Large Image Figure ViewerDownload Hi-res image Download (PPT)