Incidence, risk factors, and consequences of radiographic pin migration after pinning of pediatric supracondylar humeral fractures

Abstract

Current literature on pin migration is inconsistent and its significance is not understood. We aimed to investigate the incidence, magnitude, predictors, and consequences of radiographic pin migration after pediatric supracondylar humeral fractures (SCHF). We retrospectively reviewed pediatric patients treated with reduction and pinning of SCHF at our institution. Baseline and clinical data were collected. Pin migration was assessed by measuring the change in distance between pin tip and humeral cortex on sequential radiographs. Factors associated with pin migration and loss of reduction (LOR) were assessed. Six hundred forty-eight patients and 1506 pins were included; 21%, 5%, and 1% of patients had pin migration ≥5 mm, ≥10 mm, and ≥20 mm respectively. Mean migration in symptomatic patients was 20 mm compared to a migration of 5 mm in all patients with non-negligible migration (P < 0.001). Pin migration > 10 mm was strongly associated with LOR [odds ratio (OR) = 6.91; confidence interval (CI), 2.70–17.68]. Factors associated with increased migration included increased days to pin removal (β = 0.022; CI, 0.002–0.043), migration outwards versus inwards ( = 1.02; CI, 0.21–1.80), and BMI > 95th percentile (OR = 1.63; [1.06–2.50]). Factors not associated with migration included cross-pinning, number of pins, and fracture grade. In summary, we identified a 5% incidence of radiographic pin migration ≥ 10 mm and determined the factors associated with it. Pin migration became radiographically significant at >10 mm where it was strongly associated with LOR. Our findings contribute to the understanding of pin migration and suggest that interventions targeting pin migration may decrease the risk of LOR. Level of Evidence: Level III – Retrospective Cohort Study

Incidence, risk factors, and consequences of radiographic pin migration after pinning of pediatric supracondylar humeral fractures

“…The above classifications [2][3][4][5] picture only transverse fractures, not considering that there are different fracture patterns, as described by Bahk et al [6] in the Johns Hopkins classification system, with medial/lateral and sagittal oblique patterns resulting in increased shear forces across the fracture site, possibly requiring more pins. Bahk et al [6] stated that fractures with a higher degree of obliquity are more prone to rotational and/or extension malunion, stressed the importance of stability testing and reported 3 technical observations, with none of these aspects having been investigated by Badin et al [1], despite the patients having been treated in the same department and the senior author of both publications having been the same.…”

“…Badin et al [1] identified a significant association between 'days to pin removal' and magnitude of pin migration. However, the effect was NS in the univariate analysis (UVA) but was still included in the multivariate analysis (MVA), despite it being common practice only to include those variables in the MVA which were identified as significant in the UVA [7].…”

“…However, the effect was NS in the univariate analysis (UVA) but was still included in the multivariate analysis (MVA), despite it being common practice only to include those variables in the MVA which were identified as significant in the UVA [7]. Badin et al [1] also included a large number of independent variables in the MVA which can create collinearity (two or more independent variables are closely related to one another, undermining the statistical significance) and can result in overfitting. The authors [1] pointed out that the significant associations between 'height' and 'days to pin removal' and 'magnitude of pin migration' were clinically irrelevant so it would possibly have been better to exclude these from the statistical model.…”

“…Badin et al [1] also included a large number of independent variables in the MVA which can create collinearity (two or more independent variables are closely related to one another, undermining the statistical significance) and can result in overfitting. The authors [1] pointed out that the significant associations between 'height' and 'days to pin removal' and 'magnitude of pin migration' were clinically irrelevant so it would possibly have been better to exclude these from the statistical model.…”

“…However, it has been recently identified that this definition of MLOR is invalid [9], possibly underestimating LOR by the factor of about 4. Therefore, Baldin et al [1] have possibly hugely underreported LORs. The authors did not provide information on the quality of the intra-operative reductions and pin positioning.…”