Is rod diameter associated with the rate of rod fracture in patients treated with magnetically controlled growing rods?

Abstract

Few risk factors for fracture in magnetically controlled growing rods (MCGR) have been identified. We hypothesize an increased rate of rod fracture in small diameter rods compared to large diameter rods in patients with early-onset scoliosis (EOS). The purpose of this study was to determine the association between the diameter of MCGR constructs and the rate of rod fracture. Patients with EOS who underwent MCGR implantation-primary or conversion-from 2013 to 2018 were identified from two registries including 40 centers. Rod diameter sizes greater than 5.0mm or less than or equal to 5.0mm were defined as "Large" and "Small" rods, respectively. Only dual-rod constructs were included. The primary outcome measure collected was rod fracture at any point in treatment up to the most recent follow-up. Cox regression was utilized for unequal follow-up to compare rate of breakage at the last follow-up between cohorts. 527 patients with 1,054 rods were included. 552 (52.4%) rods had a diameter of less than or equal to 5.0mm and 461 (43.7%) rods had a diameter of greater than 5.0mm. 41 (3.9%) rods were missing a recorded rod diameter and were not included in the analysis to determine the association between the rate of fracture and rod diameter. 20 (1.9%) total rod fractures occurred: 9 (1.6%) rods with diameters of ≤ 5.0mm, 10 (2.2%) rods with diameters of > 5.0mm, and 1 uncategorized rod (p = 0.529). No difference in the rate of rod fracture or survival distribution was found between rod diameters of > 5.0mm and ≤ 5.0mm even after stratification by ambulatory status, major coronal curve, weight, or location of anchors. Rod fracture appears to be a rare event in dual MCGR constructs and rod diameter does not seem to be associated with the incidence or rate of rod fracture. Surgeons may consider other criteria for selecting rod diameter in their patients such as patient size, amount of surgical correction, single vs. dual constructs, and risk of hardware prominence.