After obtaining local Institutional Review Board approval (registration number 899,534), we retrospectively studied the patients with high grade (> 50% slippage) lumbar spondylolisthesis that was treated with an intraoperative temporary distraction rod at our Orthopedic Department in Imam Reza Hospital, Mashhad, Iran, from August 2009 to May 2012. We only included the patients with high grade spondylolisthesis who were clinically symptomatic and refractory to aggressive conservative treatment for more than three months, although in skeletally immature patients there was a high probability of slip progression; asymptomatic patients with high grade slippage were also the surgical candidates. The patients who suffered from trauma, infection, or tumor-induced spondylolisthesis, the patients with significant underlying generalized diseases (like Paget disease, osteogenesis imperfecta, severe osteoporosis, etc.), or the patients with less than two years of follow-up were excluded from this study. We also excluded very immature or anomalous patients in whom safe insertion of pedicle screws could not be guaranteed.

All of the mature patients signed the informed consent form. For the patients who had not reached the legal age of consent, the informed consent forms were signed by their legal guardians. Preoperatively, magnetic resonance imaging scans, standing anteroposterior and lateral radiographs were taken for all of the patients. Slip angle was measured as the angle formed by the L5 inferior endplate and the line perpendicular to the posterior aspect of S1.3) Slip percentage was calculated as a percentage of forward displacement of the upper vertebra on the top of the lower vertebra.1) Correction rate was measured by dividing the difference in slip percentage (pre- and postoperative slip-page) by preoperative slip percentage multiplied by 100.

We finally evaluated 21 patients (14 females, 66.7% and 7 males, 33.3%) with a mean age of 50.4 ± 9.2 years (age range, 34 to 64 years). The location of slip was L5-S1 in 19 patients (90.5%) and L4-L5 in two patients (9.5%). The mean follow-up period was 39.2 ± 11.8 months (range, 25 to 60 months). Based on the one-sample Kolmogorov-Smirnov test, our data had a normal distribution. Clinical and radiologic data of our treated patients in the pre- and postoperative period are summarized in Table 1.

We did not exclude patients with spondyloptosis; but there was no case of severe spondylolisthesis among our patients. Immediate and final correction rates showed no statistically significant difference (p = 0.470). Throughout the follow-up period, a 3.6% ± 3.7% loss of correction was observed, and this amount of recurrence had no significant impact on the total correction rate (p = 0.091). The amount of loss of correction also did not show any correlation with the duration of follow-up (r = 0.097, p = 0.677).

Intraoperative dural tear occurred in two patients, and it was repaired properly and it healed without any further complications. Due to intractable postoperative radicular pain in one patient, re-insertion of a pedicle screw became necessary and then the pain improved immediately. Postoperative worsening of the neurologic status did not occur in any of the patients. Two patients developed superficial wound infections that were cured with local wound care and oral antibiotics. Implant failure and nonunion (at least symptomatically) did not occur in any of the patients, although we did not perform computed tomography (CT) scanning routinely in all of the cases. Subjective satisfaction rate is shown in Table 2. There was no case with a poor surgical outcome and overall satisfaction (good and excellent results) was observed in 19 patients (90.5%). Although the amount of loss of correction observed throughout the follow-up was not significant, the satisfaction rate had a reverse correlation with the amount of loss of correction (r = -0.517, p = 0.16).

An assessment of the history of spine surgeries including those for spondylolisthesis indicates that most of the spine surgeons tend to perform spinal procedures via a posterior approach. The reasons that are usually cited include more familiarity with this approach, decreased risk of injury to great vessels or vital organs, greater ease of revision operations, ability to operate on multi-levels, and no need for assistance from a general or vascular surgeon.17,18,19) In surgical treatment of high grade spondylolisthesis, various approaches (posterior only, staged or simultaneous anterior and posterior approaches) have been reported in the literature, but it would be ideal to identify a familiar posterior approach that can achieve all surgical goals and provide appropriate long-term results.2,7) In this study, we think we could achieve this purpose with the use of temporary intraoperative distraction technique.

In surgical treatment of high grade spondylolisthesis, much controversy exists regarding the selection of vertebral reduction or in situ fusion among spine surgeons. It seems that due to the availability of more powerful and smaller implants that are more likely to achieve a successful fusion and provide better clinical satisfaction, reduction seems to be more popular.20) Hence, a variety of reduction techniques and tactics have been invented by numerous authors. Many of them have raised the issue of gradual reduction of the slipped vertebra.21,22) Karampalis et al.21) used Magerl's external fixator for this purpose and then performed circumferential fusion in 9 patients with high grade spondylolisthesis and measured both clinical and radiological outcomes after about 11 years. The improvement in slip magnitude (Meyerding classification), slip angle, lumbosacral angle, sacral rotation, and sacral inclination were 2.9 grades, 66%, 47%, 51%, and 47%, respectively. Although the patients' expectations were met, solid fusion was achieved in 88.9% of the patients. Mean postoperative ODI and low back outcome scores were 8% and 56.6%, respectively. The author finally recommended this technique as a powerful and safe technique without any neurologic complications that may be associated with other procedures. In comparison, the number of patients in our study was more than twice the number of patients in the above mentioned study and we had one patient with radicular pain who needed revision surgery for reinsertion of a pedicle screw. The mean improvement in slip magnitude and postoperative ODI in our study were 1.5 grades and 8.2%, respectively.

On the other hand, in 1996, Hu et al.5) reported the surgical results in 16 cases with high grade spondylolisthesis that was treated with posterior decompression and reduction using the Edwards Modular Spine System. This technique could decrease slip percentage from 89% to 29% and slip angle from 50° to 24°, respectively. Three cases developed a neurologic deficit and 4 patients had implant failure. Excellent, good, and fair results were obtained in 10, 5, and 1 patient, respectively. In conclusion, the authors themselves acknowledged that it was a complex technique and it may be associated with significant neurologic or hardware complications. In our study, the technique could improve slip percentage from 59% to 21% and slip angle from -8° to -15°, respectively. The comparison shows that the severity of vertebral slippage in our study was somewhat lesser. In our study, there was only one case (4.8%) of neurologic complication and implant failure did not occur in any of the cases, but the overall surgical satisfaction was similar.

In a study of 34 patients with high grade L5-S1 spondylolisthesis, Lengert et al.4) evaluated spondylolisthesis reduction and maintenance over time with L4-S1 (25 patients) versus L5-S1 (9 patients) fusion. They used a lever-arm system and posterior fusion in combination with lumbosacral graft. In their patients, the mean Taillard spondylolisthesis index decreased from 64% to 37% and L5-S1 segmental lordosis increased from 11° to 18°. They observed a loss of reduction from 19° to 14° in the L5-S1 group, while the reduction was maintained in the L4-S1 group. These authors finally concluded that this system provided satisfactory reduction of spondylolisthesis, but in high grade L5-S1 spondylolisthesis, loss of reduction was better controlled with posterior L4-S1 fusion versus shorter L5-S1 fusion. In our study, we routinely included one upper intact vertebra in instrumentation to reduce the probability of implant failure and significant loss of correction; therefore, we could not compare the results of short fusion versus long fusion but our results showed that long fusion was associated with non-significant loss of correction.

Our study had some limitations. It had a retrospective design and it inevitably had the inherent limitations of retrospective studies. The number of patients was not significantly large, although it was not too small. CT scanning was not performed routinely in all of the cases for assessing the fusion rate, and this certainly reduced the accuracy of fusion assessment. We suggest that in the future, a larger prospective study should be conducted by different surgeons using a similar technique and the results should be assessed by a neutral third person. In conclusion, in surgical treatment of refractory high grade spondylolisthesis, the use of a temporary distraction rod to reduce the slipped vertebra in combination with neural decompression, posterolateral fusion, and longer instrumentation is associated with satisfactory clinical and radiologic outcomes and this method can be the method of choice for surgical management of these difficult cases.