Healing the humeral shaft nonunion: Prior surgery confers increased risk of recalcitrant nonunion

OBJECTIVES: To determine the rate of successful humeral shaft nonunion repair in patients with no prior surgery on the humerus (failed nonoperative management), compared to patients with a history of prior surgery on the humerus (initial operative treatment complicated by nonunion, or prior attempted nonunion repair after failed nonoperative management). METHODS: Design: Retrospective Setting: Two academic trauma centers (one level 1 and one level 2) Patient Selection Criteria: All skeletally mature patients undergoing nonunion repair of a presumed aseptic humeral shaft nonunion (AO/OTA 11A, 11B, 11C, 12A, 12B, 12C) were eligible for inclusion. Outcome Measures and Comparisons: The primary outcome was osseous union. Univariate analysis was used to examine patient, injury, and treatment factors associated with recalcitrant nonunion between those with and without prior surgery. RESULTS: One hundred fifty-nine patients were included. Eighty-two patients had a history of prior operative treatment. The group with prior operative treatment was significantly younger (47 vs 52, p=0.047) and had fewer comorbidities (average Charlson comorbidity score 1.3 vs 1.9, p=0.015). There were 34 men in the group with prior operative treatment, compared to 37 in the group without (p=0.493). For patients with prior operative treatment, 17/82 (21%) developed a recalcitrant nonunion, versus 2/79 (3%) in patients with no prior operative treatment (p<0.001). The number of prior operations on the arm was significantly associated with increased risk of recalcitrant nonunion (3% risk if no prior surgeries, 19% risk with one prior surgery, 25% risk with 2 prior surgeries, 33% risk with 3 prior surgeries, p=0.004). No demographic factors were associated with development of a recalcitrant nonunion (p>0.05 for all). Nine patients had unexpected positive cultures but this was not associated with increased risk of recalcitrant nonunion (22% in patients with infection vs 26% in those without, p=0.907). CONCLUSIONS: Patients undergoing nonunion repair after prior operative treatment of a humeral shaft fracture had a 1 in 5 rate of recalcitrant nonunion, while patients undergoing initial nonunion repair after failed nonoperative management had a 3 in 100 rate of recalcitrant nonunion. Increased risk of persistent nonunion stemmed not from initial treatment strategy for the acute fracture, but rather from the presence of any prior surgery.