Introduction Implant instability has a relatively low rate of incidence, however revision interventions in those cases are expensive, difficult to perform and have a considerable frequency of poor results. Even though various novel technologies are available, this problem is not resolved yet and it is essential to assess risk factors that may lead to this complication. The objective of this study is to determine risk factors for the pedicle screw instability development after spinal instrumentations. The objective of this study is to assess risk factors in relation to pedicle screw instability development. Materials and Methods This is a prospective randomized study of 120 spinal instrumentations performed in case of traumatic injuries and spinal degenerative diseases of lumbar and thoracolumbar spine. Preoperatively patients underwent CT examination and bone density was measured in Hounsfield units. CT scans utilized slice thickness 0.5 mm. Tube voltage was 120 kV, current 300 mA, auto mA range 180 to 400; 1.0 s/3.0 mm/0.5 × 32, and helical-pitch 21.0. Transpedicular fixation was used to treat patients either as standalone technique or in combination with various types of interbody fusion, also if indicated, decompression of nerve roots and spinal cord was performed. Patients with suboptimal screws placement were excluded from this study. The duration of follow-up was 1 year; cases with screws loosening were registered. Logistic regression analysis was utilized to assess the relationship between expected risk factors (extension of decompression, bone density, extension of fixation), and the rate of screws loosening. Results Extensive decompression resulting in a total or subtotal bilateral facet joints total resection results in the considerable increase in frequency of pedicle screws instability. The parameters of the logistic regression model were chi-square = 14.78227, p = 0.0001209; В0 = −0.9360933, р = 0.0005580619; В1 = 1.479709, р = 0.0002964492; OR (odds ratio) = 4.391667 (2.001693; 9.635212). It was estimated that the rate of implant instability has a strong inverse relation to the bone density measured in HU during CT investigation. The parameters of the logit regression model were chi-square = 12.32050, p = 0.0004486; B0 = −1.846886, p = 0.007698078; B1 = 0.01737968, p = 0.00169618; OR = 56.7059 (4.706; 683.2748). Even though the extension of fixation in general turned out to be irrelevant to a screw loosening, it was pointed out that S1 vertebra included into a fixed area is an additional risk factor for implant instability formation. The estimated parameters of logistic regression model were chi-square = 4.882991 p = 0.0271296; B0 = −0.5221894, p = 0.01585754; B1 = 0.992193, p = 0.03159398; OR = 2.697143 (1.093079; 6.655126). The general nonlinear regression model that includes all listed previously contributing factors has a predictive value of 78.33%, goodness of fit chi-square = 34.96233 p = 0.0000001. Conclusions Apparently, surgical intervention should be planned thoroughly avoiding destabilizing overextended decompression because this could be a significant contributing factor for the implant instability formation in future. Bone quality must be taken in view while planning spinal instrumentation. If necessary, augmentation with bone cement must be performed to prevent implant instability formation. The extension of fixation to the level S1 must be clearly justified as it was estimated that it is an additional considerable risk factor for pedicle screws loosening.
Read full abstract