In this study, a retrospective cohort and a case-control analysis was conducted to investigate whether opportunistically assessed BMD was associated with the risk of reoperation following instrumented LSF with and without PMMA-augmentation. The retrospective cohort analysis showed that opportunistic BMD assessment on perioperative CT scans allows detection of osteoporotic bone density in patients who are scheduled for LSF. The prevalence of osteoporosis in this elderly group of patients undergoing elective surgery is rather high. We were able to demonstrate that patients with PMMA-augmented LSF exhibited much lower BMD than those with non-augmented LSF, while both groups showed an almost equal rate of reoperations. The case-control analysis showed that lumbar BMD was similar in patients with reoperation compared to matched controls who did not have reoperation. Although not being statistically significant, BMD was numerically lower in cases with reoperations than in controls without reoperation.
Results of this work are in line with previous studies. Low BMD assessed using HU measurements on preoperative CT scans has been associated with adjacent vertebral body fractures after spinal fusion (15). In a case-control study, decreased HU on preoperative CT scans were associated with symptomatic pseudarthrosis on one-year follow-up after posterolateral lumbar fusion (37). Patients with radiographic signs of screw loosening and non-fusion on follow-up after posterior lumbar interbody fusion with pedicle screw fixation had significantly lower areal BMD assessed by DXA compared to patients without these signs (14). Moreover, patients with signs of screw loosening in CT following posterior spinal fixation showed significantly lower BMD assessed by opportunistic screening than patients without signs of loosening (22). Of note, decreased BMD in the cervical spine was identified as the major predisposing factor for the occurrence of traumatic odontoid fractures in elderly patients (32).
Given that patients in our retrospective cohort analysis with PMMA-augmented LSF had significantly lower BMD than those without augmentation, it is noteworthy that these patients did not show a higher reoperation rate. PMMA-augmented screw fixation is recommended in osteoporotic bone (4,5), because it improves the fixation and fatigue strength ex vivo (38), reduces the risk of screw loosening and pullout (39), and increases fusion rates with maintained correction angles in vivo (40). In reviews, PMMA-augmented screw fixations and other technical modifications like long-segment constructs is considered to reduce the risk of instrumentation failure in osteoporotic patients (4,41,42). Multiple points of fixation have been recommended in the osteoporotic spine for a long time (43). To avoid ending within a spinal transition zone or a kyphotic section long-segment constructs seems to be beneficial (25,44,45), as these regions are typically prone to adjacent segment degeneration, adjacent vertebral body fractures, or implant failure. In the presented study, patients with augmented LSF had also one more fused segments on average. There might be a positive effect of both these factors – augmentation and longer constructs – leading to a slightly numerically decreased reoperation rate (12% vs. 15%), despite substantially lower BMD. However, a larger scale study is needed to verify if a BMD below 80 mg/cm³ (threshold for osteoporosis) is suitable as a decision point for the use of augmentation, particularly considering other factors such as number of fused levels.
In the retrospective cohort analysis, patients with elective LSF had a mean age of 67 (range 30–89) years and showed a prevalence of 52% for osteoporosis. Women had slightly, but non-significantly lower BMD than men. There was a predominance of women in the group with PMMA-augmented LSF (79%) showing substantially lower BMD compared to non-augmented LSF, which is probably due to postmenopausal changes in bone metabolism. When only looking at the patients who received non-augmented LSF, the prevalence of osteoporotic BMD was still high with 29% vs. 85% in the augmented group. The high prevalence of osteoporosis may be due to the high mean age of our study population. However, indications for LSF in elderly patients persist and favorable clinical outcome after surgery can be achieved in the majority of these patients regardless of an increased overall surgical risk (46). Previously, a prevalence of about 30% of osteoporotic BMD or fragile bone strength has been reported in women between the age of 50 to 70 years undergoing spinal fusion (8). However, equally high rates of osteoporotic BMD were observed in the population of men and women older than 50 years of age prior to spinal fusion surgery (7,9). Thus, biomechanical considerations and the use of the aforementioned surgical techniques are of increasing importance when performing spinal instrumentations in the osteoporotic spine (42,45).
Of note, BMD measurements are not performed on a regular basis prior to surgery in our hospital. We performed opportunistic screening of lumbar BMD in clinical CT scans (31), which has been validated and applied in various studies (22,28–30,32–34,47), showing low short- and long-term reproducibility errors (30). Previous studies demonstrated that lumbar BMD can be assessed in sagittal reformations of contrast-enhanced MDCT and used to differentiate patients with and without osteoporotic fractures (30), and it could also be used to predict these fractures (29). Linear correction equations can be adjusted for systematic bias of apparent bone density related to different calibration techniques and contrast application (48). We analyzed CT scans obtained on different MDCT devices for indications other than densitometry and applied asynchronous calibration to calculate BMD (22,34). In contrast to direct HU measurements, which are dependent on the specific MDCT device used, calibration enables inter-scanner and/or inter-study comparability of BMD values. Moreover, predefined BMD thresholds for osteoporosis can be used (36). In order to determine whether there is an increased risk of complications and reoperation after LSF, we estimated a BMD threshold of 83.7 mg/cm³ for patients with non-augmented surgeries in our study. In a biomechanical study, BMD of less than 80 mg/cm³ was associated with early screw loosening and unsatisfactory spinal fixation (19). Specifically, a previous study hypothesized that an areal BMD below 0.674 ± 0.104 g/cm² indicates a potentially increased risk of spinal fusion failure (14). Although difficult to compare to volumetric BMD, this value certainly lies within the osteoporotic range. Apparently, the estimated cutoff in our study matches closely with the proposed threshold of lumbar BMD for the diagnosis of osteoporosis (36). This emphasizes the high importance of assessing lumbar BMD prior to spinal instrumentation in the neurosurgical setting.
The presented analyses are based on retrospective data collection and, therefore, are prone to bias. Loss of follow-up is a major confounding factor in the retrospective analysis approach, though expected to be similar across groups. Moreover, there are no objective criteria for reoperation. The decision to have revision surgery is inherently subjective and influenced by the patient’s and neurosurgeon’s preferences, amongst other factors. Furthermore, the authors are aware that sagittal balance of the spine is another important biomechanical factor, which can have an impact on the outcome subsequent to LSF. Unfortunately, long-standing radiographs that would have allowed the analysis of the sagittal vertebral axis before and after surgery were not available for the presented data since they were not part of routine perioperative workup in our institution regarding the time of data inclusion.