BackgroundCurrently, there remains a lack of consensus regarding factors predictive of complication such as re-nonunion after primary revision in femoral shaft nonunion subsequent to failed intramedullary nailing (IMN). A better understanding of prognostic factors could potentially reduce the risk of re-nonunion happening and allow patients to maximize their recovery in the most expeditious manner. Our study aims to identify risk factors in the development of re-nonunion after primary revision inclusive of exchanging reamed nailing (ERN) and augmentative compression plating (ACP) with IMN in situ for femoral shaft nonunion subsequent to failed IMN.MethodsA retrospective study was performed for 63 cases (61 patients) of femoral shaft nonunion subsequent to failed IMN, who were made primary revision with either ERN or ACP from June 2007 to June 2015. The following set of variables was selected based on the speculation that they would contribute to the outcome: sex (male or female), age, body mass index(BMI), smoking, alcohol abuse, cause of injury, fracture type, type of IMN (antegrade or retrograde), use of IMN locking screws(dynamic or static), site of nonunion, primary nonunion time, pathological type of nonunion, bone defect (mm), primary revision method (ERN or ACP), and adjuvant autogenous bone grafting (ABG) (yes or no). Univariate analysis and multiple regression were used to identify risk factors in the development of re-nonunion after primary revision with either ERN or ACP for femoral shaft nonunion subsequent to failed IMN. The minimum follow-up time was 1.5 years (standard deviation [SD] = 1.2, range 1.5–8 years).ResultsOf 63 cases (61 patients) of femoral shaft nonunion subsequent to failed IMN, primary revision with ERN was performed in 33 (52.4%) cases and primary revision with ACP was performed in 30 (47.6%) cases. Adjuvant ABG procedure was undertaken in 39 (61.9%) cases during primary revisions. Re-nonunion was diagnosed as in 18 (28.6%) cases after primary revision with either ERN or ACP. There was a significant difference in time to union between patients treated with primary ERN and those with primary ACP (log-rank, p = 0.006). Furthermore, the difference was also statistically significant between patients with adjuvant ABG procedure and those without it (log-rank, p = 0.009). The relative risk factors included smoking, BMI, site of nonunion, bone defect, primary revision method, and adjuvant ABG procedure. However, primary revision method and adjuvant ABG procedure were shown to be two independent risk factors in multiple logistic regression analysis.ConclusionsPatients with excessive tobacco use, BMI ≥ 30 kg/m2, bone defect ≥ 5 mm, primary revision with ERN, and no adjuvant ABG procedure had a higher likelihood of developing re-nonunion. Of these risk factors, primary revision with ERN and no adjuvant ABG procedure were two strongest risk factors.