Relationship of the character of rod fractures on outcomes following long thoracolumbar fusion to the sacrum for adult spinal deformity

Abstract

BACKGROUND CONTEXTRod fractures (RF) and pseudarthrosis are a frequent occurrence after adult spinal deformity (ASD) surgery and may be problematic. However, not all RF signal nonunion and cause clinical concern. An improved understanding of the sequelae after RF occurrence is valuable for further management. PURPOSETo characterize the radiographic findings, clinical outcomes, and revision rates between patients who developed unilateral RF (URF) and bilateral RF (BRF) following thoracolumbar posterior spinal fusions to the sacrum for ASD and identify patient characteristics associated with clinically significant RF that lead to subsequent revision surgeries and detection of nonunion. STUDY DESIGN/SETTINGA retrospective single-center cohort study was performed. PATIENT SAMPLEPatients undergoing long-construct posterior spinal fusions to the sacrum performed at a single institution from 2004 to 2014 and developed a RF postoperatively were included. OUTCOME MEASURESPatient demographics, radiographic parameters, surgical data, Oswestry Disability Index (ODI), Scoliosis Research Society-22 (SRS-22), and revision rates. METHODSInclusion criteria were ASD patients age >18 who had ≥5 vertebrae instrumented and fused posteriorly to the sacrum and development of RF. Data were compared among patients: who developed unilateral-nondisplaced RF (UNRF), unilateral-displaced RF (UDRF), bilateral-nondisplaced RF and bilateral-displaced RF (BDRF) at baseline and follow-up. ODI and SRS-22 scores were assessed at baseline, 1 year postoperatively, the time of RF occurrence, and latest follow-up. RESULTSOf 526 patients who met inclusion criteria, 96 (18.3%) developed RF (URF n=70 [73%]; BRF n=26 [27%]). Preoperative demographics and surgical parameters were similar between the groups. BRF patients had substantial loss of sagittal correction from 1-year postoperatively to the time of RF, including loss of sagittal vertical axis (4.8 cm vs. 2.2 cm; p<.001), loss of lumbar lordosis (14.8° vs. 4.9°; p=.010) and loss of pelvic incidence minus lumbar lordosis mismatch (PI-LL) mismatch (5.0° vs. 14.6°; p=.020) compared with those of URF patients. The BDRF group had more loss of ODI scores (13.4 vs. 4.2; p=.013), SRS pain score (0.8 vs. 0.2; p=.024), SRS function score (0.3 vs. 0; p=.020) and SRS subscore (0.4 vs. 0.1; p=.148) from 1-year postoperatively to the time of RF and underwent revision surgery more often than the UNRF group (87.5% vs. 4.8%; p<.0001). At final follow-up (median 2.8 years, range 1–10.3 years after RF detection), URF patients who did not undergo revision surgeries still maintained equivalent sagittal alignment correction (sagittal vertical axis, LL and PI-LL; all p>.05) and had similar, not worse, mean ODI scores, SRS Subscore and SRS pain compared with the time at RF and 1-year follow-up. CONCLUSIONSRF are not uncommon after ASD operations. Asymptomatic, UNRF in our study did not jeopardize clinical outcomes or radiographic alignment parameters and, in most cases, did not represent a nonunion, as opposed to BRF. BRF patients exhibited loss of sagittal correction, loss of clinical outcome improvements, as measured by ODI, SRS pain and SRS Subscore at the time of RF, and were revised more often than URF patients.