Comparison between topping-off technology and posterior lumbar interbody fusion in the treatment of chronic low back pain

Supplemental Digital Content is available in the text


Introduction
Posterior lumbar interbody fusion (PLIF) has been considered to be the standard surgical treatment for patients suffering from chronic low back pain (CLBP) caused by lumbar degenerative disease. [1] In addition to its satisfactory clinical outcomes, there are still a series of complications in PLIF. [2] The increase in the range of motion (ROM) of adjacent segments will cause the acceleration of adjacent segment diseases (ASDs), which are the most commonly seen complications in the follow-up. [3] Posterior dynamic stabilization systems, including interspinous process device (IPD); pedicle-based stabilization devices (PDS); total facet replacement system, have been used to decrease ASDs after PLIF since 1980 s. [4] The most commonly used hybrid dynamic stabilization system is "topping-off " technique. This technique combines the rigid fusion with dynamic nonfusion of adjacent segments, such as IPD or PDS, in order to reduce the hypermobility and overstress of the adjacent intervertebral disks. [5][6][7][8][9] At the present study, there have been various flexible systems in spinal motion preservation technology, including PDS systems and IPS systems. The most commonly used PDS systems include Dynesys, [10] NFlex, [11] Isobar TTL, [12] CD Horizon [5] and DSS. [13] Furthermore, the widely used IPS systems include Coflex, [14] DIAM, [15] Wallis. [16] Because there is a lack of clear clinical evidences, the difference between topping-off technique and PLIF in postoperative outcomes is still controversial. [9,17,18] Therefore, a meta-analysis was carried out in this study to compare all available data on postoperative clinical and radiographic outcomes of topping-off technique and PLIF in the treatment of CLBP.

Search strategy
The present review was conducted according to preferred reporting items for systematic reviews and meta-analyses statement. [19]

Inclusion criteria and exclusion criteria
The inclusion criteria included: clinical evaluation was followed up for no less than 1 year; conservative treatment was frustrated on the treatment of CLBP; PLIF or topping-off surgery was used in the treatment; patients were in the same preoperative radiographic baseline. Exclusion criteria included: biomechanical studies and non -human or in vitro studies; abstracts, case reports, expert opinions, and non-comparative studies; therapies for tumors, infections, revision surgeries or congenital malformations.

Data extraction and quality assessment
All data was extracted from the text, pictures, and tables of the articles by 2 authors (SYS, JZD). The data included: age, gender, duration, study design, enrolled number, radiographical adjacent segment disease (RASD) and clinical adjacent segment disease (CASD), ROM, global lumbar lordosis (GLL), visual analog scale (VAS) of back and leg (VAS-B, VAS-L), Oswestry disability index (ODI), Japanese Orthopaedic Association (JOA) score, duration of surgery, estimated blood loss (EBL), reoperation rates, complication rates. If there is a disagreement about the outcomes, other authors (WW, SBL) would participate in the discussion to reach a consensus. Newcastle-Ottawa scale (NOS) [20] was used to assess the quality of the included studies by 2 authors (XYS, CK). A full score of 9 stars and a score of 7 or more is an excellent quality study.

Data analysis
The statistical analysis of the studies was performed in RevMan5.3 software. Odds ratios (OR) and standardized mean difference (SMD) with 95% confidence intervals (CI) were used for dichotomous data and continuous data. Heterogeneity across trials were explored according to the results of Chi-squared test and I 2 statistic. Random-effects model was considered if there was a significant heterogeneity assumed as P value less than .05 and I 2 > 50%. Otherwise, data were pooled by using the fixedeffects model. If there was a potential heterogeneity, subgroup analysis and sensitivity tests would be performed in conjunction with possible clinical realities. Publication bias was analyzed by the funnel plot.
ROMs of upper intervertebral levels were discussed in two articles. [14,22] The fixed effect model was applied considering  Table 1 Characteristics of included studies.   there is no significant heterogeneity between two groups (I 2 = 51%). No significant between-group difference was found in this analysis (SMD -0.36; 95% CI -0.89, 0.17; I 2 = 0%; P = .71, Fig. 5)

Discussion
It is still controversial that whether ASD is a natural evolution of an aging spine or a consequence of spinal fusion. [28,29] Previous studies stated that the risk factors for ASD included age over 50 years, [30,31] sagittal imbalance, [32] increased length of fusion and surgical approaches. [33] The elastic fixation in topping-off technique can act as a partially active buffer between fused segments and proximal mobile segments. Therefore, the ASD could be reduced by topping-off technique via the application of dynamic fixation, which could prevent the proximal adjacent segments from degenerating. [34] Because there is still a lack of uniformity in the criteria of RASD, its definition in the previous studies were summarized: spondylolisthesis increase more than 3 mm, loss of intervertebral disc height, dynamic angulation of the interspinous space less than 10°. [7,14,15,21,24,25,35] Considering that there is no significant heterogeneity in the analysis of RASD and CASD in our results, our research has good reference value. Our study showed that the incidences of proximal RASD and CASD in the PLIF group were significantly higher than those in topping-off group. The results of this study are consistent with previous studies. [9,36] However, the result of distal RASD rate showed no significant betweengroup difference. This implied that topping-off technique was more effective in preventing proximal ASD. The possible explanation is that the compensatory mechanism of ROM in the proximal segments in the topping-off group can be improved. [34,37] Our results showed that No significant between-group difference was found in ROMs of upper intervertebral levels. Therefore, the decrease of compensatory mechanism in topping-off technique may work through multiple segments. The postoperative GLL was similar between topping-off group and PLIF group in our results. This indicates that the expansion effect in topping-off technology can be negligible in global spine compared with PLIF. However, the correction of GLL may not be an advantage of topping-off technology. [23,34] The recovery of lumbar functions and the curative effects of operation were evaluated by JOA score, ODI and VAS. Our results showed that VAS-B in the topping-off group was significantly less than that in the PLIF group; however, no significant between-group difference was found in VAS-L. This  showed that topping-off technique was more effective in relieving CLBP. However, JOA score and ODI were similar between both groups. The possible explanation is that scoring systems are associated with the increasing age and the complications of dynamic equipment. [38] VAS gives more weight to subjective feelings of patents, while JOA score and ODI are more focused on the objective motor functions of patients. [25,26] All of these indicate that topping-off technique, compared with PLIF, is more effective in improving subjective feelings of patents rather than objective motor functions. This conclusion is in consistent with the founding of previous studies, which showed that topping-off technique could achieve a good clinical improvement even in the long-term follow-up. [4,39,40] Our study showed that no significant difference was found in EBL and duration of surgery between topping-off group and PLIF. The possible explanation is that additional exposure of anatomical structures is not needed in the insertion of dynamic implants; this will save the operation time. [4] Considering there are many confounding factors in the included studies, this result needs to be interpreted carefully.
Whether topping-off technique can decrease the incidence of complications after fusion surgery or not has been inconclusive. The most commonly seen postoperative complications in toppingoff technique are screw loosening, screw fracture and spinous process avulsion fracture. [15,21,23,27] Screw loosening is most likely to be found in Hybrid Stabilization Devices; in addition, spinous process fracture is the most commonly seen complication in Interspinous Process Devices. [41,42] The rates of complications, such as dural tear, infection, implant loosening, pseudoarthrosis, and implant breakage, were discussed in our study. However, no significant difference between topping-off group and PLIF group was found in our results. The "halo zone" in dynamic stabilization systems show that the forces conveyed from the dynamic implant can increase the stress on rigid fixation over time, and then implantassociated adverse events will occur. [38] However, our study indicated that, compared with PLIF, the application of the toppingoff technique would not be influenced by this effect. The application of hydroxyapatite coated pedicle screws might be an ideal method to prevent implant related complications. [43] Previous studies showed that hydroxyapatite could promote bone deposition on the implant surface and promote the formation of direct chemical bonds between the implant and the bone interface, which might reduce the complication rates in patients. [28,44] It has been reported that ASD after PLIF is not a complication caused by the fusion itself; ASD is more likely to be cause by normal aging process. [45] However, Lu et al. [15] reported that different methods of surgical treatments had greater influences on ASD than age and gender. Similarly, our results showed that no significant between-group difference was found in age and gender. This indicated that age and gender have a relatively smaller influence on the outcomes of the surgery.
There are several limitations in this meta-analysis. First, lumbar degeneration was a series of diseases in which the overall outcome could vary depending on specific diagnosis, such as intervertebral disc herniation, stenosis, and spondylolisthesis. Second, most of the included studies were retrospective studies. As a result, there would be an inherent limitation associated with the risks of reporting or selection bias. Third, different types of dynamic devices used in adjacent segments could have different structures, which might influence the outcomes. Therefore, more randomized controlled trials were needed to verify the results of this study.

Conclusions
Topping-off can effectively prevent the ASD from progressing after lumbar internal fixation, which is be more effective in proximal segments. Compared with PLIF, topping-off technique was more effective in improving subjective feelings of patents rather than objective motor functions. However, no significant difference between topping-off technique and PLIF can be found in the rates of complications.