Outcomes of internal and external fixation in midfoot Charcot neuroarthropathy: a systematic review and single-arm meta-analysis

Abstract Background Midfoot Charcot neuroarthropathy (CN) reconstruction is often performed using internal fixation (IF) or external fixation (EF). This systematic review summarizes clinical outcomes of these techniques including fusion, amputation, return to ambulation and complication profiles. Methods A PRISMA-guided systematic review identified studies of midfoot CN reconstruction using IF or EF. Random-effects single-arm meta-analyses of logit-transformed proportions were used to pool postoperative amputation and fusion; fixation type was explored using mixed-effects meta-regression. Robustness was evaluated using leave-one-out analyses, influence diagnostics and patient-limited reweighting for bilateral reconstructions. Small-study effects were assessed with Egger’s regression. Results Seventeen observational studies were included. The pooled postoperative amputation rate was 8.7% (95% CI 5.8–12.9; PI 3.6–19.7; k = 15), with meta-regression showing no significant association with fixation type ( p = 0.163). Subgroup estimates were 10.2% after IF (95% CI 5.6–17.7) and 6.4% after EF (95% CI 5.1–7.9). Fusion pooled at 80.2% (95% CI 59.9–91.7; PI 17.5–98.7; k = 11), with substantial heterogeneity not explained by fixation type ( p = 0.631). Subgroup pooling yielded fusion estimates of 81.5% (95% CI 59.8–92.9; k = 9) for IF and 75.6% (95% CI 0–100; k = 2) for EF, noting imprecision for EF due to limited studies. Return to ambulation was 94.1% in IF and 81.3% in EF. Infection, hardware failure, and revision rates were 21.8%, 31.6%, and 31.7% in IF, and 5.7%, 12.5%, and 19.1% in EF, respectively. Patient-limited reweighting produced near-identical pooled estimates (amputation 8.9%; fusion 79.9%). Conclusion Midfoot CN reconstruction generally achieves high limb salvage and fusion overall. When stratified by fixation category, ambulation and fusion were higher in IF, whereas amputation, infection, hardware failure, and revision rates were also higher. Between-category differences may reflect confounding by indication and should be interpreted as hypothesis-generating. These findings support patient-tailored fixation selection and suggest that differences largely reflect multiple patient factors rather than a causal effect of fixation type. Fixation choice is also likely to be influenced, to some extent, by surgeon expertise and familiarity with specific reconstructive techniques.