Results of instrumented posterolateral fusion in failed back surgery

Abstract

The success rate following revision surgery for failed back surgery is inherently low. However, a general consensus for a satisfactory surgical intervention after failed disc surgery has not yet been reached. We present a cohort study of 25 adult cases treated by instrumented posterolateral fusion with or without nerve root exploration/decompression for recurrence or persistence of symptoms after previous disc surgery. The average age at surgery was 47 years (range 39–56 years). Complications were seen in four patients in the form of intraoperative dural tears. A satisfactory clinical outcome was seen in 80% of the patients, and the overall fusion rate was 93%.

Introduction

Failed back surgery is a nonspecific term implying that the final outcome of surgery did not meet the expectations of both the patient and the surgeon that were established prior to surgery [15]. It is a common puzzle that has confounded physicians as much as patients over the history of spinal surgery [3].

The most common structural causes of failed back surgery are foraminal stenosis, painful disc, pseudoarthrosis, neuropathic pain, recurrent disc herniation, facet pain, sacroiliac joint pain, and iatrogenic instability [4, 16].

A thorough investigation of the cause is essential. Magnetic resonance imaging (MRI) is the mainstay investigation. The addition of gadolinium will enhance in cases of vascular epidural fibrosis; the opposite is true in cases of nonvascular recurrent disc herniation [2, 18]. The role of discography is controversial because it is believed to be very subjective; and, the results must be interpreted cautiously in patients with failed back surgery [15].

The success rate in revision spinal surgery is inherently low [6], and the treatment depends on the cause. In revision disc surgery, unless the cause is pure recurrence of disc herniation or pure nerve root involvement, the addition of fusion is felt to be mandatory [3] because the occurrence of micro-instability following disc surgery has been found to be inevitable [9–11, 14].

There are many ways to carry out a spinal fusion, but simpler methods carry the fewest complications [10]. It is received wisdom that simultaneous instrumentation carries a greater chance of achieving a fusion, and pedicular fixation is the technique most commonly used [5].

Our operative approach involved posterolateral fusion with pedicular fixation. The posterior approach to the spine allows for a simultaneous discectomy and/or nerve root decompression when necessary.

The purpose of this study was to review the outcome of adopting the above rationale in treating patients with persistent, recurrent back or back and leg symptoms after previous disc surgery.

Patients and methods

We conducted a prospective study of 25 consecutive adult patients between 1998 and 2002.

Our inclusion criteria included patients with failed disc surgery who presented with back or back and leg pain (17 complained of back and leg pain, and eight complained of predominant back pain). Three of the 25 patients had unilateral paraesthesia in the distribution of the L5 dermatome following the primary operative treatment. Patients were considered for surgery if they had little improvement, with persistence or recurrence of back and/or leg pain following the previous spinal operations despite conservative treatment for a minimum period of 6 months. All were adjudged to have a significant reduction in their quality of life. Patients were excluded from the study if they were involved in medicolegal claims and if the underlying pathology was diagnosed to be pure recurrent disc herniation (at the same or different level), lateral recess stenosis, or discitis.

The group consisted of 15 females and ten males. The average age at surgery was 47 years (range 39–56 years). The average duration of symptoms prior to surgery was 28 months (range 17–38 months).

Preoperatively all patients were evaluated with plain X-rays, including dynamic flexion and extension views to detect the presence or absence of signs of instability including traction spurs, reduction in disc space height, and translation. All patients had gadolinium-enhanced MRI scans of their lumbar spine.

Previous operations included 12 at L5-S1 (eight with previous formal laminectomy and four with previous hemilaminectomy), nine at L4-5 (three with previous laminectomy and six with previous hemilaminectomy), three double-level at L4-5/L5-S1 (two with previous formal laminectomies at the two levels and one with hemilaminectomies at the two levels), and one patient had a three-level laminectomy at the L3-4/L4-5/L5-S1 levels (Table 1).

Table 1.

Previous preoperative procedures

	Formal laminectomies	Hemilaminectomies
Total	18	12
L3-4	1	0
L4-5	6	7
L5-S1	11	5

A total of 30 levels (n=30) were fused using a routine instrumented posterolateral fusion: one-level fusion in 21 patients(n=21), two levels in three patients(n=6), and three levels in one patient (n=3). Patients with back and leg pain were subjected to an additional re-decompression of the nerve root, but no attempt was made to excise any perineural fibrosis if encountered.

All patients were followed up clinically and were assessed using the criteria reported in 1995 by Ricciardi et al. [13]. This involves assessing the presence or absence of symptoms, the use of analgesics, the patient's level of function, and radiological evidence of fusion. On this basis, the postoperative results were graded as poor, fair, good, or excellent. The radiological work-up for follow-up included anteroposterior and lateral lumbosacral spinal X-rays, and fusion was judged to be solid if there was a continues sheet of uninterrupted bone trabeculae. In cases of doubtful fusion, computed tomography (CT) was used to assess fusion. All radiological investigations were analysed by independent viewers (M.A.E. and Y.A.R.).

Patients were seen routinely at 3, 6, 9, and 12 months and then yearly following surgery. The minimum postoperative follow-up was 2 years. The mean follow-up was 3 years (range 2–5 years).

Results

The mean operating time was 3 h (range 2–5 h), and the mean blood loss was 950 ml (range 500–2,300 ml). The average hospital stay was 8 days (range 5–16 days).

Using the Ricciardi criteria [13], 15 cases were graded as excellent, five as good, three as fair, and two as poor. Considering excellent and good results to be satisfactory, 20 cases (80%) were adjudged to have a satisfactory clinical and radiological outcome. Out of the 30 levels where fusion had been attempted, 28 (93%) were adjudged to be solidly fused (CT scanning was used to confirm fusion in six levels). Two levels (7%) were adjudged to show non-union; each was a single-level fusion.

There were four (13%) intraoperative minor to moderate dural tears that were repaired by suturing. There were no infections and no graft-related problems.

Discussion

Segmental instability is the key factor in the decision-making process when treating patients with failed disc surgery. Structures that provide spinal mobility are also responsible for preventing translation, and their degrees of contribution are as follows: intact facet capsule, 39%; intact disc and annulus, 29%; supraspinous and intraspinous ligaments, 19%; and ligamentum flavum, 13% [1]. Even interventions at the hemilamina and the ligamentum flavum may change both load-bearing and kinematic characteristics of the spine, leading to hypermobility, accelerated bone degeneration [9, 12], irritation of receptors related to the facet joints [12], and lower back pain leading to mechanical instability. Changes in the spine's load-bearing and kinematic characteristics may also lead to mechanical deformation of the intraspinal nerve tissues, inducing nerve root pain [9, 10] and leading to neurological instability.

The precise diagnosis of instability may present a clinical challenge and depends on radiological as well as clinical clues, including instability and painful catch [11], translation of three millimetres or more on flexion and extension radiographs [8], traction spurs, and decrease of the disc height more than 30%. In this study, 23 of our patients had at least one clinical and one radiological criterion of instability.

Sunny et al. reported an 8% incidence of dural tears in their study [17]. In our study, nerve root exploration was undertaken in all cases with leg pain, particularly in cases with equivocal MRI scan, which might explain the relatively high incidence of dural tears (13%) and raises a question about the exact indication of re-exploration of nerve roots in revision spinal surgery.

In addition to its role in mechanical stabilisation of the motion segment, spinal fusion can also reduce radiculopathy by creating local traction on the tethered nerve root in the direction of the segmental motion [17].

The overall result achieved in this study was satisfactory (excellent or good) in 20 patients (80%); they all showed solid radiological union (Figs. 1a,b, 2, 3a,b). Five patients (8%) had unsatisfactory (fair or poor) outcomes. Two of these five patients were the cases adjudged not to have united, and confirmed by CT scanning. Clinically, these two patients presented with unsatisfactory levels of lower back pain postoperatively, and therefore it was felt that a second revision was justified. The two patients underwent anterior revision surgery using tricortical iliac graft when progression of fusion was felt to have come to a halt. At the final follow-up, solid anterior fusion had been achieved and the clinical outcome was good.

Fig.  1.

a, b Preoperative anteroposterior and lateral radiographs

Fig. 2.

Preoperative magnetic resonance image

Fig. 3.

a, b 2-year follow-up anteroposterior and lateral radiographs showing solid fusion

Three patients claimed a less satisfactory outcome despite radiological solid fusion. At the end of the follow-up period, they reported persistent paraesthesia in the distribution of the L5 dermatome. The persistence of paraesthesia was felt to be the cause of the unsatisfactory outcome because there were no signs of radiological non-union, and the improvement of back pain was satisfactory. The patients were subjected to nerve root stimulation as well as passive modalities of treatment with no improvement. It was judged that there would be no benefit from re-exploration of the neural elements because that might provoke more epidural adhesions and thereby worsen paraesthesia.

The clinical and radiological results achieved in this study compare favourably with the best reported in the literature [7, 19].

Conclusions

We believe the relatively good results achieved in this study can be attributed in part to proper patient selection, as all the patients included in the study had a history of a previous single failed disc surgery with no history of longstanding chronic neuropathic pain. We believe that mechanical segmental instability is inevitable following disc surgery and therefore plays an important role in the pathogenesis of failed disc surgery. Accordingly, unless the recurrence of symptoms is due to pure recurrent disc herniation, spinal fusion is mandatory in the treatment of failed disc surgery. Routine nerve root exploration is not indicated in most cases, particularly if root involvement can be excluded by gadolinium-enhanced MRI scans.