Background
Transforaminal lumbar interbody fusion (TLIF) is commonly used to treat patients with lumbar degenerative disease [
1]. TLIF offers the advantage of obtaining a circumferential arthrodesis through a unilateral approach with minimal retraction of neural elements. Many surgeons have reported good clinical and radiological outcomes for such lumbar interbody fusion techniques. [
2‐
5] However, there are some patients who have post-operative residual, lower extremity pain, or numbness even without post-operative complications [
2‐
5].
Loss of lordosis has been thought to be one of the factors influencing post-operative residual symptoms [
6,
7]. In 2010, Schwab et al., reported the importance of maintaining global sagittal alignment when performing long-segment spinal fusion surgery in the treatment of adult spinal deformity [
8]. Their results clearly showed that post-operative global sagittal balance is significantly correlated with post-operative residual pain and disability [
8]. Radiological parameters that most highly correlate with pain and disability are sagittal vertical axis (SVA), pelvic tilt (PT), and the balance of pelvic incidence (PI) and lumbar lordosis (LL) [
9]. Schwab et al. proposed radiological parameter thresholds predictive of worse clinical symptoms and poorer quality of life, and concluded that a PI-LL (PI minus LL) of 10° or less was the ideal spinopelvic alignment for reducing post-operative pain and disability [
9]. Therefore, surgeons should plan corrective surgeries to acquire an adequate LL that achieves a harmonious spinopelvic alignment (PI-LL 10° or less).
In previous studies [
8‐
10], the importance of maintaining appropriate PI-LL has been emphasized, particularly for long-segment fusion for adult spinal deformity; however, there are few studies evaluating the influence of PI-LL on surgical outcomes following short-segment fusion. In this study, we examined the effect of PI-LL mismatch on surgical outcomes of short-segment TLIF for lumbar degenerative diseases.
Methods
Patients with lumbar degenerative diseases, such as lumbar spinal stenosis, degenerative spondylolisthesis and spondylolytic spondylolisthesis, treated by short-segment 1 or 2 level TLIF were included in the present study (Table
1). Patients who showed any indication of other pathological problems, such as vertebral fractures, tumors, or infectious diseases, were excluded. Fifty-two consecutive patients meeting the criteria were divided into Group A (PI-LL ≤ 10°:
n = 22) and Group B (PI-LL ≥ 11°;
n = 30) by post-operative radiological evaluation.
Table 1
Fusion levels and indications for fusion in groups A and B
Number of fusion levels | | 1.1 ± 0.4 | 1.4 ± 0.5a |
Level of fusion | (No. of cases) | | |
One level | L3-L4 | 0 | 1 |
| L4-L5 | 16 | 12 |
| L5-S1 | 4 | 5 |
Two levels | L3-L4-L5 | 1 | 9 |
| L4-L5-S1 | 1 | 3 |
Indications for fusion | | | |
Degenerative spondylolisthesis | 14 | 19 |
Spondylolytic spondylolisthesis | 2 | 1 |
Spinal canal stenosis | 6 | 10 |
Pre-operative characteristics of patients, including age, sex, number of fusion levels, body mass index (BMI), and incidence of degenerative scoliosis, diabetes mellitus, and depression, were compared between the two groups.
Anteroposterior and lateral radiographs in the neutral standing position were taken pre-and post-operatively. On anteroposterior radiographs, pre-operative scoliotic curvature was evaluated using the Cobb method. On lateral radiographs, PI and LL were evaluated one-year after surgery. The PI is measured as the angle between a line drawn perpendicular to the sacral end plate at its midpoint and a line drawn from the midpoint of the sacral end plate to the midpoint of the femoral head axis. The LL is the sagittal Cobb angle measured between the superior end plate of L1 and the superior end plate of S1 [
8]. Post-operative parameters, such as PI, LL, and PI-LL, were then compared between the two groups. Excellent inter- and intra-rater reliability has been reported for these measurements [
10].
Evaluation of surgical outcomes and clinical symptoms
Intra-and post-operative complications were retrospectively reviewed for each group. Surgical outcomes were evaluated using the Nakai score (Table
2: 3 = excellent to 0 = poor) [
11]. Pre-and post-operative patient symptoms were also assessed by the visual analogue scale (VAS: scores 0-100 mm; for LBP, lower-extremity pain, and lower-extremity numbness) and the Oswestry disability index (ODI). In addition, the detailed VAS scoring system, originally developed by us for LBP while in motion, standing, and sitting, was used for a more detailed evaluation of LBP, pre-and post-operatively [
12]. Post-operative data were prospectively acquired for at least one year following surgery. All scores were compared between the two groups.
The patient has resumed work-related and other activities with slight or no symptoms. | 3: Excellent |
The patient has resumed work-related and other activities but occasionally feels pain in the back or lower limbs after strenuous work. | 2: Good |
The patient has reduced work-related and other activities because of residual pain in the back or lower limbs. | 1: Fair |
The patient cannot work or carry out activities of daily living and is considered to be disabled. | 0: Poor |
Surgical procedures
All patients underwent surgery in the prone position after induction of general endotracheal anesthesia. The side selected for the unilateral approach was the side showing more neurological involvement. If symptoms were equal on both sides, the approach was made on the left side. A midline incision exposed the posterior elements and the lateral aspect of the facet joint. Following ipsilateral laminotomy, partial facetectomy exposed the surface of the intervertebral disc. Resection of the bone was performed manually using an osteotome and Kerrison rongeur to preserve local bone. Following the removal of disc material and meticulous endplate preparation, the resected local bone was milled and packed into the intervertebral disc space as a bone graft. One or two fusion cages packed with milled local bone were then inserted into the disc space. Contralateral decompression was performed if needed; the contralateral ligamentum flavum, ventral aspect of the lamina, and medial aspect of the contralateral facet joints were resected through the ipsilateral side. Following neural decompression, pedicle screws were placed on both sides. A 1-cm fascia incision allowed insertion of pedicle screws transfascially, using a percutaneous pedicle screw system. Adequate compressive force by the pedicle screws was applied to the disc space to establish stability of the fusion cages.
The percutaneus pedicle screws used for the patients in this series were Viper (Depuy Synthes Spine, Raynham, MA) or Mantis (Stryker, Kalamazoo, MI).
Written informed consent was obtained prior to surgery. The study protocol was approved by the institutional ethics committee of Toho University Sakura Medical Center (No. 2012–071).
Statistical analysis
Statistical analyses used were the unpaired t-test for age, fusion levels, BMI, PI, LL, PI-LL and VAS, the Mann-Whitney U test for the Nakai score and ODI, and the chi-square test for sex and incidence of each disease.
Multiple regression analysis was performed to investigate the relative influences of pre-and post-operative PI-LL on each pre-and post-operative VAS, detailed VAS, and ODI after adjustment for age, sex, fusion levels, BMI, presence of scoliosis, diabetes mellitus and depression. PI-LL was utilized as the independent variable and pre-and post-operative data for VAS and ODI as the dependent variables.
Probability values of p < 0.05 were considered significant. Values are expressed as the mean ± standard deviation (SD).
Discussion
Post-operative flat-back syndrome has been reported to be a factor causing undesirable surgical outcomes after lumbar fusion surgery [
6,
7]. Schwab et al. recently proposed a systematic evaluation method to classify sagittal spinal alignment in flat-back cases [
8]. Their study demonstrated that SVA, PT, and PI-LL were most closely related to LBP and disability. Since the Scoliosis Research Society-Schwab adult spinal deformity classification was established [
10], there have been no studies examining the influence of PI-LL mismatch on surgical outcomes after short-segment lumbar interbody fusion. Our study is the first to show that PI-LL mismatch significantly correlated with surgical outcomes after short-segment lumbar interbody fusion. Consequently, our study reveals that surgical outcomes after short-segment lumbar interbody fusion will be poorer when good spinopelvic balance is not achieved following surgery. Thus, surgeons should pay attention to the sagittal spinopelvic balance of patients even when treating patients with short-segment lumbar spinal stenosis or spondylolisthesis. Our results show that post-operative PI-LL mismatch causes greater residual LBP. Interestingly, in this study, our originally developed detailed VAS system was successful in revealing that LBP while standing is most closely correlated with post-operative PI-LL mismatch. Patients with flat-back syndrome are usually characterized by an inability to stand erect and back pain [
7]. In our study, post-operative residual LBP while standing may reflect the inability to stand erect due to flat-back syndrome.
We believe that PI-LL did not correlate with pre-operative symptoms (Table
6) because pre-operative symptoms, including pain caused by nerve root compression, spinal instability, and spinopelvic alignment, are more complex than post-operative symptoms. Post-operative symptoms, particularly LBP while standing, were more simply influenced by spinopelvic alignment (Table
7), because symptoms caused by nerve root compression and spinal instability may have been relieved by fusion surgery. Therefore, PI-LL significantly correlated with post-operative LBP, but not with pre-operative LBP (Tables
6,
7).
There are some limitations to this study. Most importantly, post-operative residual symptoms are affected by multiple factors. In our study, Group A showed patient characteristics that differed from those of Group B. Although most were not significant (Tables
1,
3), the possibility that differing characteristics influenced the results of the comparison study between Group A and Group B persists. In particular, the difference in number of fusion levels may affect surgical outcomes because better surgical outcomes would be expected from single-level fusion than two-level fusion [
13]. However, the multivariate analysis adjusted for age, sex, fusion levels, BMI, presence of scoliosis, diabetes mellitus and depression showed that PI-LL was significantly associated with post-operative symptoms, not with pre-operative symptoms. These results lead us to conclude that post-operative PI-LL mismatch causes more intense post-operative residual symptoms, particularly LBP while standing. There are other limitations, including the short post-operative follow-up and small number of patients, and the variety of pathologies that were indicators for fusion surgery in the study. There is a concern that short-term outcomes may be influenced by surgery-related factors, rather than radiologic factors. However, in our study no patients with surgical complication had deleterious consequences from the complication after additional treatments, such as screw reposition, hematoma removal, or dural suture were accomplished. Recent studies reported that a post-operative PI-LL mismatch resulted in increased load on adjacent segments and predisposition to adjacent segment pathologies and late surgical complications seen after longer follow-ups [
14‐
17]. Thus, longer follow-up times may reveal an effect of adjacent segment pathology or late surgical complications on surgical outcomes. However, the statistically significant differences found between our two study groups, despite the small number of subjects, lead us to believe our results are reliable and provide important information for spine surgeons treating lumbar degenerative disease. Further investigations in which inclusion criteria are stricter with a longer follow-up and larger number of subjects are needed to confirm our conclusions.
In this study, we did not examine the effects of correcting LL on surgical outcomes because the surgery in this series was not intended to correct LL. Therefore, we cannot conclude that the surgical correction of LL would have a favorable effect on surgical outcomes. However, there have been several studies showing the positive effects of improving sagittal spinopelvic alignment on surgical outcomes of short-segment fusion surgery. Kim et al. found that patients with improved PT after posterior lumbar interbody fusion (PLIF) had better clinical results [
18]. Lazennec et al. reported that a larger post-operative PT positively correlated with LBP after lumbosacral fusion surgery [
19]. A larger PT, which reflects pelvic retroversion, is usually correlated with loss of LL [
20]; therefore, we surmise that the surgical correction of LL decreases PT and produces improved surgical outcomes. Hioki et al. found that two-level PLIF provided satisfactory results by preserving lordosis of the lumbar spine, suggesting the importance of restoring LL by fusion surgery [
21]. However, Lazennec et al. reported that post-operative loss of lordosis was not always related to post-operative residual symptoms [
19]. One explanation for this inconsistency may be that the appropriate LL varies depending on a patient’s PI. These findings indicate the importance of acquiring ideal PI-LL when performing short-segment lumbar fusion.
More recently, Kim et al. reported that TLIF has the potential to cause post-operative flat back, and that PLIF is more reliable than TLIF for restoring spinopelvic alignment [
22]. For the maintenance of segmental lumbar lordosis, extreme lateral interbody fusion (XLIF) is thought to have more capability to increase disc height and correct spinopelvic alignment [
23‐
25]. From these observations, surgeons can select the appropriate surgical technique from the various fusion methods, TLIF, PLIF, XLIF, multi-segment fusion, spinal osteotomy, etc., taking into consideration the ability of each method to restore spinopelvic alignment. Efforts should be made to reduce PI-LL to 10° or less whenever feasible.
Competing interests
The authors declare that they have no competing interests.
Author’s contributions
Author contributions to the manuscript preparation include the following. YA and AN drafted the manuscript, HT, MS, FT, and MS participated in interpretation and acquisition of data, KT and SO participated in its design, conception, and helped to draft the manuscript. AW, TN, MT and SO collected patients’ information and participated in co-ordination and co-wrote the manuscript. KN participated in its design and statistical analysis. YE participated in data acquisition and manuscript revision. All authors read and approved the final manuscript.