Morcellized local grafts as cost effective option for interbody fusion in thoracolumbar fracture dislocation: Seven years follow up of 53 patients

Abstract

Introduction

Traumatic fracture dislocation of the spine injury is essentially a three column injury that optimally needs surgical intervention to decompress, stabilize and fuse the spinal column. This study evaluate the outcome of posterior and posterolateral decompression, instrumentation and 360 degree fusion achieved with help of locally harvested autologus morcellized grafts in traumatic fracture dislocation of thoracolumbar spine.

Methods

53 patients were included in this retrospective study. Patients aged 16-55 years, single level fracture dislocation of thoraco-lumbar spine (D5-L5) were included. Patients with multiple level fractures, coexisting degenerative diseases of spine ,pathological fractures, patients presenting more than three weeks after initial trauma, patients with concomitant severe head injury that necessitated emergency surgery for the same were excluded from the study. Patients underwent posterior and posterolateral decompression , posterior instrumentation and interbody as well as posterolateral fusion with use of morcellized bone from resected posterior elements. Follow up data at immediate post operative period ,12 months and yearly thereafter upto minimum 7 years was obtained from previous record.

Results

There were 46 males and 7 females. Mean age was 31.15 ±9.64 yrs .Mean follow up period was 7.4 yrs(range 7-10 yrs). Thoracolumbar dislocation was most frequently noted at thoraco lumbar junction (T10-L2).Thirty six patients had complete neurological deficit (ASIA A) and sixteen had incomplete neurology. At one year follow up, osseous fusion was noted in 48 (90.56%) patients and 5 patients (9.44%) had fibrous union which was determined on CT scan. Immediete post operative, one year and 7 year kyphosis angle was calculated and change in kyphosis angle was not statistically significant. There was no implant failure till last follow up.

Conclusion

Morcellized locally harvested autologus grafts are sufficient to achieve 360 0 spinal fusion in fracture dislocation of thoracolumbar spine.

1. Introduction

Traumatic fracture dislocations of the spine, essentially a three column injury constitute less than 3% of thoracolumbar spine trauma¹ that necessitates surgical intervention to decompress, stabilize and fuse the spinal column. There are several approaches for surgical fixation of such fractures like anterior, posterior and combined anterior and posterior approach. Long segment fixation (at least two segments above and below the injured vertebra) is commonly employed to stabilize the spinal column in fracture dislocation cases.² An interbody fusion at dislocation level combined with poster lateral fusion minimizes the stress on implants ensures a long term satisfactory outcome.

There is paucity of literature that demonstrates long term outcomes of relatively short segment fixation (2 segments above and one segment below with index level pedicle screw application) in fracture dislocation.^3,4 Spinal fusion is usually achieved with interbody cage in modern practice. There is also limited literature showing the usefulness of morcellized local grafts as a sole agent to obtain interbody fusion.⁵ Moreover long term follow up results of such grafts to maintain the spinal alignment are still lacking. The primary aim of the study is to evaluate the efficacy local grafts obtained from resected posterior elements to obtain interbody fusion in long term follow up of seven years. The secondary aim of the study is to evaluate the outcome of relatively short segment fixation in fracture dislocation patients in the long term.

2. Material and methods

Retrospective analysis of 64 patients was performed after approval from institutional review board for fracture dislocation of thoracolumbar spine from Jan 2010 to Dec 2013. Five patients died within 1 year of surgery due to associated chest and abdominal injury and six patients were lost to follow up. Hence, 53 patients were included in this retrospective study. Patients with multiple level fractures, coexisting degenerative diseases of spine, pathological fractures, patients presenting more than three weeks after initial trauma, patients with concomitant severe head injury that necessitated emergency surgery for the head injury was excluded from the study.(Table 1). The patients were stabilized initially with standard ATLS protocols. Radiographs and CT scan and MRI of the whole spine was done subsequently. Standard anticoagulation protocol was followed. Surgery was performed once patient was stabilized and radiographic evaluation was complete. Patients were mobilized with TLSO brace, day after surgery for three months and thereafter without the brace. Mobilization protocol was same regardless the fixation construct in all the patients in our study. Patients were followed up clinically and radiologically immedietly after surgery; at 3,6 and 12 months and yearly thereafter. At each follow up, radiograph of thoraco-lumbar spine- AP and lateral views were taken and a CT scan was done at 12 months post surgery, to assess fusion status. Assessment of fusion was performed with serial radiograph post operatively at 3,6,12 months and yearly thereafter and compared with immediate post operative radiograph. Thin slice helical CT with sagittal and coronal reconstruction was performed to assess the fusion mass at 12 months (The radiographic criteria we used to assess the maturity of fusion mass is summarized at Table 2). Each patient meeting all the criteria described in “Definite” group was classified as “osseous fusion” otherwise the patient was classified as “Indeterminate/Doubtful” fusion status. To minimize radiation hazards of CT scan we did not perform serial CT scans unless reoperation was planned. The patients whose radiographic features at 1 year follow up indicated as “Indeterminate/Doubtful” fusion were kept under regular follow up with radiographs. A repeat CT scan was performed if significant implant motion was detected after one year follow up radiograph or presence of progressive deformity was confirmed (more than 5° change of cobb's angle in comparison to immediate post operative radiograph).

Table 1.

Inclusion and exclusion criteria of patients included in the study.

Inclusion criteria	Exclusion criteria
1.Single level fracture dislocation	1.Multiple level fracture dislocation
2.Patients presenting within three weeks of initial trauma	2.Coexisting degenerative diseases of spine
	3.Pathological fractures
	4.Patients presenting more than three weeks after initial trauma
	5.Patients with concomitant head injury

Table 2.

Radiographic criteria for assessment of fusion mass.(Each patient must satisfy all the criteria described in “Definite” group to be classified as osseous fusion otherwise the patient is classified as “Indeterminate/Doubtful” fusion status.).

	Radiographic features (plain radiographa and CT scanb)
Definite (osseous fusion)	1.Presence of bridging trabecular bone across the vertebral end plates.2(for interbody fusion) 2.Continuous bony density with absence of complex radiolucencies across the transverse process.3(for posterolateral fusion) 3.Criteria mentioned in “indeterminate/doubtful” are absent.
Indeterminate/Doubtful (non osseous fusion)	1.Graft resorption12 2.Implant subsidence or migration12 3.No definitive bridging trabeculae across graft host interface, linear lucency between the fused vertebral body22 4.Cystic changes across the vertebral end plates.22 5.Presence of intradiscal gas, adjacent medullary sclerosis and endplate cystic changes.42(secondary signs of micromotion at fusion site intervertebral region). 6.Increase of kyphosis angle more than 5° at dislocation level in comparison to immediate post operative radiograph.

^aPlain radiograph of each follow up was compared with immediate post opearative radiograph to assess fusion.

^bFine-cut helical CT with sagittal and coronal reconstruction was performed to assess the fusion mass.

We assessed fusion with CT scan at one year and then followed the patients of “osseous fusion” in out patient clinic every one year. Patients were advised to visit to hospital after six months if fusion was classified as “non osseous” at one year post operative CT scan.

Segmental hypnotic angle was measured from radiograph obtained immediately after surgery, at 12 months post operative and at final follow up at or after 7 years using cob's method (angle between upper end plate of normal vertebra superjacent and lower end plate of normal vertebra infrajacent to the fractured level). Database was searched and immediate post operative, 12 months and final follow up (7 years) data was retrieved and analyzed.

3. Operative procedure

Patients were taken up for surgery on an urgent basis because of the highly unstable nature of injury. Under general anaesthesia, patient was positioned prone on a radiolucent table, with standard posterior midline exposure. In 48 cases of our series, proximal segment had translated anteriorly. In these cases, pedicle screws were inserted two levels superior and inferior to the dislocation level. A temporary rod was secured in the distal segment pedicle screws first and initial distractive force was applied. After the initial distraction, rod reducers were employed to translate the proximal segments posteriorly to achieve the desired reduction by gradually tightening the rod reducers on both sides alternatively. In cases with posterior translation of proximal segments (5 cases in our series), reduction pedicle screws were implanted in distal segments and the temporary rod was secured to proximal segment. Reduction was achieved by alternate gradual tightening of set screws of the reduction screws in the distal segment. An initial attempt was made at preservation of facet joints. However, in cases with difficult reduction partial or complete facetectomy was performed.

Adequate decompression of neural elements was routinely performed in all cases. A complete laminectomy was done routinely and the spinal canal was explored for compressive elements like retropulsed bone fragments, crushed intervertebral discs etc. Bilateral facetectomy was performed if required, to aid reduction, as well as to access the intervertebral disc space bilaterally using posteromedial approach. Intervertebral disc space was cleared off all disc material and end plates were prepared for interbody bone grafting and fusion. Morcellized graft was prepared from the locally harvested bone obtained from decompression and used to achieve 360° fusion (posterior, interbody and posteroateral). In most of the cases, the morcellized grafts were placed through unilateral posterolateral interbody approach in an already distracted spinal segment using temporary rod.

All the patients were started in bed mobilization next post operative day and wheelchair mobilization in brace within a week after surgery.

4. Results

There were 46 males and 7 females with mean age of 31.15 ± 9.64 years. Twenty four (24) patients had history of fall from height, twenty two (22) patients reported history of road traffic accident and seven cases (7) were due to fall of heavy objects on back. Mean follow up period was 7.4 yrs (range 7–10 yrs). The demographic details of patients included in the study is summarized in Table 3.

Table 3.

Demographic characteristics of patients included in the study.

1.Number of patients	64
2.Completed follow up	53
3.Lost to follow up	6
4.Death during follow up	5
5.Number of pts included in study	53
Age
1.Mean ± SD	31.15 ± 9.64
2.Range
Sex
1.Male	46
2.Female	7
Mode of injury
1.Fall from height	24
2.Road traffic accident	22
3.Fall of heavy object on back	7

Thoracolumbar dislocation occurred mostly in thoraco lumbar junction T10-L2. All the fractures were rotationally unstable type C as per AO classification. (The detailed involvement of vertebra level frequency is summarized in Fig. 1.) In most of the cases, we noted anterior translation of proximal spinal segment relative to distal column (43 cases). Posterior and lateral translation of proximal vertebral column was noted in five cases each. Detailed neurological assessment was done as per ASIA scoring system at each follow up. Pre operatively 36 pts had ASIA A, 14 were ASIA B and 2 pts were ASIA C and one pt was ASIA E. At one year follow up at, 34 pts were ASIA A, 10were ASIA B,5 were ASIA C,3 were ASIA D and one pt was ASIA E (Fig. 2). Two patients improved from ASIA A to ASIA B while 34 patients with ASIA A did not show any neurological improvement. There were 16 patients with incomplete neurological deficit. Neurological improvement was noted in 7 patients among them. 4 pts improved from ASIA B to ASIA C,2 patients improved from ASIA B to ASIA D and one patient improved from ASIA C to ASIA D. One patient had no neurological deficit preoperatively [7]. No further neurological improvement was noticed after one year in our patients. There were no patient who were deteriorated in post operative period. Mean time from injury to surgery was 6.58 ± 5.20 days (range 1–21 days).40 patients were operated within 10 days of injury and 13 patients were operated after 10 days of injury. In most of the cases the surgical delay is attributed to delayed visit to our hospital. Intraoperatively, dural tear was detected in 21 patients in accessible locations (39.6%) and repaired accordingly with primary suturing. Dural tears in inaccessible locations (like lateral recess around pedicle and anteriorly due to retropulsed fragments) were left alone for fibrous healing to minimize further neural damage. All of them were neurologically ASIA A and one patient improved neurologically as ASIA B postoperatively.

Fig. 1.

Frequency of vertebral level involvement of fracture dislocation.

Fig. 2.

Pre and post operative neurodeficit frequency among patients according to ASIA grade.

Postoperative complications included superficial wound infection in 6 patients of which 5 were treated with dressing and antibiotics. One patient had to undergo debridement and secondary suturing. Drain was removed in all cases after 3 days except one patient who had CSF leak in which case it was removed after 12 days.

The osseous fusion status was determined with serial radiographs and CT scan. We noted bridging bones between the vertebral bodies with matured trabecularisation in majority (90.56%) of our cases and there was no case of implant failure even after 7 years of follow up (Fig. 5). Radiolucency around eleven (11) pedicle screws was noted in six (6) cases. But as these patients were asymptomatic they were kept under close follow up. Non osseous fusion was noted in five cases (9.34%) (Results of the study are summarized in Table 4.). All the patients who were classified as “non osseous” fusion status were paraplegic (4 patients neurologically ASIA A and 1 patient was ASIA B). These patients did not experience significant pain or deterioration in functional impairment which precluded additional surgical procedure.

Fig. 5.

Radiograph depicting D12-L1 fracture dislocation with bony fusion at 12 months.

Table 4.

Summary of results.

Injury to surgery time (days)
Mean ± SD	6.58 ± 5.20
Range	1–21
Duration of follow up (years)
1.Mean ± SD	7.8 ± 0.57
2.Range	7–10
Surgical time (minute)
1.Mean ± SD	144.17 ± 12.65
2.Range	126–173
Average blood loss (ml)
1.Mean ± SD	774.53 ± 135.78
2.Range	500–1200
Pre op VAS score	8.35 ± 0.68
Post op (1 year) VAS score	1.41 ± 0.77
Fusion status
1.Osseous fusion	48 (90.56%)
2.Non osseous fusion	5 (9.34%)
Complications
A.Immediate post op
1.CSF leak	1
2.Superficial wound infections	6
B.After 1 year
1.Catheter related complications	24
2.Neuropathic pain	17
3.Pressure sore	9

We did not measure kyphosis preoperatively as severe fracture dislocation precluded the accurate measurement of angles. Immediate post operative kyphosis was 5.16 ± 2.96° and at 1 year follow up 5.89±3.17° (p 0.22) and 7 yr follow up kyphosis was 6.24±3.14° (p 0.07 comparing with immediate post op value) (Fig. 3). VAS score significantly improved at one year follow up in comparison to pre operative status (Fig. 4).

Fig. 3.

Serial measurement of kyphosis angle depicts mimimal change in the angle.

Fig. 4.

Comparison of pre op and post operative (one year) VAS score.

After one year, most common reason for hospital visit of our patients was neuropathic pain (17 cases), development of pressure sore (9 cases), urinary tract infection and catheter related complications (24 cases). Neuropathic pain responded with pharmacologic therapy based on pregabalin and amitrptyline in 16 cases but TENS therapy was employed in addition to pharmacologic treatment in one patient. The patients were evaluated periodically at pain clinic. Pressure sore was the result of inadequate nursing care in most of the patients who also suffered from malnutrition and hypoproteinemia. Regular debridement and dressing was required for five patients with grade one bed sores but required skin grafting in 3 cases and flap coverage in 1 case.

5. Discussion

Thoracolumbar fracture dislocation is a high energy injury disrupting all the three spinal columns necessitating operative intervention at the earliest possible time.⁸ The goal of the treatment is neural elements decompression and surgical stabilization of spinal column in order to rehabilitate the patient at earliest.^2,9, 10, 11 Long fixation constructs (2 segments above and below fractured level) are typically used in fracture dislocation cases to stabilize spinal column with multiple anchor points to prevent stress on hardware as well as maintain alingnment.^13,14 Recently, short segment fixation with less than 2 level above or below the fractured vertebra and fixation of the index vertebra has been used in thoracolumbar burst fracture fixation with successful outcome.¹⁵ But there is sparse literature on successful outcome with short segment fixation in fracture dislocation injuries.^16,17 We had favorable outcome in relatively short segment fixation in 22 patients along with interbody fusion (the fixation levels performed in the study is summarized in Table 5) and this is equivalent to long segment fixation both clinically and radiologically. CT scan with coronal and sagittal reformatted images were used to assess the bony fusion mass at 12 months (Fig. 6). The assessment methods of spinal fusion have been a matter of debate in recent years.⁶ Although, static and dynamic radiographs are cost effective, easily available and most commonly used method to assess fusion, their usefulness has recently been questioned due to poor reproducibility and inters observer variability. There is evidence that static as well as dynamic radiographs overestimate the fusion status^17,18,19,.²⁰ We did not obtain dynamic radiograph as most of the patients were paraplegic and obtaining a proper flexion –extension view was extremely difficult and it is often mal aligned view that precludes its use to assess fusion. Thin slice helical CT with sagittal and coronal reconstruction is an excellent and most commonly used modality for fusion assessment now a days.⁷ One year period post surgery is a reasonable time to allow fusion mass to mature and remodel. This is the reason we performed CT post one year of surgery. We did not perform repeated CT scans due to radiation hazards²¹ although few authors advocate the use of repeated CT scan at 3,6,12 and 24 months to assess fusion.¹⁶ However, if there is increase in pain or deformity or failure to achieve clinical fusion and revision surgery is planned, assessment of fusion with a repeat CT scan should be performed. We also believe that more meticulous assessment of fusion mass to be performed with CT scan in ambulatory and active patients when clinically indicated or plain radiograph is inconclusive (Fig. 7). Moreover, failure to achieve clinical fusion should be supplemented with more detailed investigations like CT scan.^5,23, 24 (see Fig. 8)

Table 5.

Fracture morphology (as per involvement of pedicles) and level of their fixation as performed in this study.

Fracture morphology (as per pedicle involvement)	Level of fixation	No. of patients in this series
1.Pedicles of the fracture-dislocation level was not intact	2 segment above and below of the dislocation level was fixed
	a.Age >35 years (transverse connector was used)	14
	b.Age <35 years (transverse connector was not used)	14
2. At least one pedicle of the fracture-dislocation level was intact	2 segment above and 1 below of the dislocation level was fixed	22
3. At least one pedicle of the fracture-dislocation level was not intact and any other pedicle within two segment above or below of the dislocation level was not intact.	Three segment above and two segment below fixation	2
4.Complete fracture dislocation with more than 10 days old injury	Complete vertebrectomy and long segment fixation	1

Fig. 6.

CT scan depicting D12-L1 fracture dislocation with bony fusion at 12 months.

Fig. 7.

Preoperative CT scan (left) with post operative CT at 12 months depicting interbody fusion.

Fig. 8.

Preoperative radiograph (left)and post operative CT scan at 12 months showing excellent interbody fusion.

Morcellized autologus graft prepared from respected posterior elements was used as the primary bone graft material in our patients. We did not use any implant for interbody fusion. Wang et al. in a retrospective study of 30 patients found similar rate of radiological interbody fusion rate with help of local grafts.⁵ Several authors have demonstrated superiority of posterior interbody fusion ²⁵that alleviates the potential complications of anterior approach.26, 27, 28 In modern practice, various types of implants are used successfully to achieve interbody fusion.29, 30, 31 These devices restore disc space, sagittal balance and distract the neuroforaminal space and maintain anatomical weight bearing within the anterior column.³² The interbody devices used commonly are in form of threaded cylindrical, rectangular or trapezoidal cages made of titanium, carbon fiber reinforced, or plain PEEK polymers, bioabsorbable polymers (i.e., polylactic acids), local auto grafts, tricortical iliac crest autografts and tricortical iliac crest allografts.33, 34, 35, 36 The disadvantages of the titanium cages are higher modulus of elasticity than bone leading to cage subsidence through vertebral end plate and inadequate visualization of fusion with radiograph as well as CT scan (due to scattering). This led to the invention of carbon fibre composites which mitigated the demerits of the titanium cages but later it was seen that there is chance of dislodgement of the implants leading to catastrophic failure.³⁷ Additionally, these implants add to surgical cost. Illiac crest autografts were previously used extensively and proved efficacious and safe but were associated with donar graft site morbidity.^38,39 Tricortical iliac rest allograft, due to its poor mechanical properties led to high failure rate.⁴⁰ We achieved interbody fusion without use of any implant but with only help of morcellized autologus interbody graft. The efficacy of spinal interbody fusion can be indirectly judged by the serial measurement of kyphosis angle. Although there was increase in kyphosis angle at one year and final follow up, the changes of the angle was not statistically significant. There was no catastrophic implant failure. These strengthen the fact that reconstruction of anterior column through the posterior approach with morcellized local graft is durable and satisfactory. Our observations support the published literature. Schmid et al. perfomed posterior lumbar interbody fusion with monocortical strut grafts in 82 cases of thoracolumbar fracture dislocation and obtained satisfactory results.²⁵ Our study represents long follow up for 7–10 years (mean 7.4 years). We observed excellent efficacy of morcellized local graft in setting of stable fixation to achieve fusion. Similar results were noted in other studies also.^2,16

There are a few distinct advantages of using local grafts to achieve fusion. The resected posterior metaphyseal cancellous elements promote satisfactory fusion while obviating any donar site morbidity. The mechanical environment is critical for successful bone healing and is closely related to the bone biology in fracture healing.⁴¹ The robust fixation of the spinal column with pedicle screw system protected the morcellized grafted bone in the intervertebral space along the axis of weight bearing, leading to bony fusion. Adequate caution be exercised in first two to three months, when the fusion mass is not strong enough and supramaximal load on implants during this time frame may cause catastrophic implant failure. Hence, in the first three months patients were rehabilitated on TLSO brace. Theoretically, there is a possibility of dislodgement of bone fragments into the spinal canal but we did not find any case with clinically significant morcellized bone graft retropulsion into spinal canal.

5.1. Limitations of the study

Randomised control trial is necessary to compare the effectiveness of local grafts to achieve fusion. In our study, there was no control group. This is a major drawback of our study. In our series most of the patients were paraplegic precluding high stress on implants. The usefulness of the local grafts to achieve interbody fusion in ambulatory active patients is yet to be determined. A properly structured biomechanical study may define the mechanical strength and efficacy of local grafts to achieve interbody fusion.

6. Conclusions

We conclude from our study that.

• 1.Morcellized local grafts are sufficient to achieve 360° spinal fusion and avoid donar site morbidity if supplemented with robust fixation in thoracolumbar fracture dislocation.
• 2.Short segment instrumentation with one or both pedicle screw fixation at index fracture level confers sufficient spinal stability provided the 360° fusion is achieved in non ambulatory patients.

Funding

No financial support was received for preparation of this manuscript.

Author Contributions

Kamran Farooque: Conceptualization, Supervision, Writing – review & editing, Formal analysis.

Vijay Sharma: Conceptualization, Supervision, Writing – review & editing.

Santanu Kar: Data curation, Writing – original draft, Formal analysis.

Declaration of competing interest

Authors declare no conflict of interest.