A Longitudinal Cohort Investigation of the Development of Symptomatic Adjacent Level Compression Fractures Following Balloonassisted Kyphoplasty in a Series of 726 Patients

Background: Balloon-assisted kyphoplasty (BAK) is a well-accepted treatment for symptomaticvertebral compression fractures (VCF) secondary to osteoporosis. Some have raised a concern ofan increased incidence of adjacent fractures due to alterations in spine biomechanics after cementaugmentation. The incidence of subsequent VCFs following BAK is poorly understood. The aim ofthis study was to investigate the timing, location, and incidence of new VCFs following BAK and toidentify risk factors associated specifically with the occurrence of new adjacent level fractures.Objectives: The study was performed to determine the incidence of symptomatic subsequentadjacent and remote level compression fractures in a cohort of patients undergoing BAK.Study Design: Longitudinal cohort investigation at an academic medical center and a centralreferral center for VCFs.Setting: A consecutive single surgeon series of 726 patients with osteoporotic compression fractures.Methods: A prospectively collected cohort of 726 patients who underwent BAK between 2001 and2014 for osteoporotic VCFs was evaluated. Seventy-seven patients were identified who underwenta second BAK for a new compression fracture and were include in the present series. The indicationfor BAK treatment was pain unresponsive to non-surgical management for all cases. Variables wererecorded for each patient, including the time between index and subsequent fracture, fracture level, andnumber of initial fractures as well as with tobacco use, body mass index (BMI), and chronic steroid use.Results: Seventy-seven of 726 patients (10.6%) underwent a second BAK procedure on average350 days following the initial procedure (range 21 to 2,691 days). Third and fourth procedures wereless common, treated in 11 and 3 patients, respectively. Forty-eight of 77 patients (62%) suffered afracture at a level immediately adjacent to the index level at mean time of 256 days. Remote levelfractures were treated at a mean time of 489 days, but no statistical difference was noted. There wasno statistically significant difference between tobacco use, BMI, and chronic steroid use betweenpatients suffering from remote and adjacent level VCFs.Limitations: This was not a population based study, and the true incidence of subsequent fracturesafter BAK might be underestimated by this analysis.Conclusions: Symptomatic compression fractures after BAK are relatively uncommon and may occurlong after the initial kyphoplasty procedure. Only half of subsequent fractures occur immediately adjacentto the initially treated level; the others occur remotely. Patients with a single symptomatic thoracic orlumbar fracture suffered from remote and adjacent level fractures equally. In contrast, all patients whosuffered both a thoracic and lumbar fracture at the same time had a second fracture at an adjacent level.Specific risk factors for remote versus adjacent level fractures could not be determined.Key words: Balloon kyphoplasty, cement augmentation, osteoporosis, vertebral compressionfracture, adjacent level fracture, vertebroplasty