The present study was conducted at a Tertiary Care Centre in New Delhi, India, between April 2012 and April 2014 on 31 patients with burst type of thoracolumbar fractures. Patients above the age of 18 years with or without neurological deficit, Kyphosis >300, anterior vertebral height loss >50%, spinal canal narrowing >40%-50%, involvement of two or more columns were included in the study. Pathological fractures and undisplaced fractures requiring conservative management were excluded from the study.

All the patients were randomly selected using the closed envelope technique. The patients were asked to open the closed opaque envelope just prior to the surgery. A thorough clinical examination was done for all the patients as per the American Spinal Injury Association (ASIA) scale [19] and Frankel grading was done to assess the severity of spinal injury. Preoperative haematological and radiological workup was done for all the patients enrolled in the study. The patients were divided into two groups. Group A (n=15) had patients in which posterior short-segment pedicle screw fixation was done while patients with pedicle screw fixation combined with posterolateral fusion were included in Group B (n=16). Well written informed consent was obtained from all the patients enrolled in the study. Ethical Committee Approval was obtained prior to the commencement of the study (DDU/00002657/2012/04). The final outcome was measured using the Modified Mcnab’s questionnaire [20], low back outcome scale of Greenough and Fraser [21] and Frankel scoring system at a regular interval of 3,6 and 12 months postoperatively. The modified Mcnab’s questionnaire is subjective which is reported between excellent and poor depending upon the severity of the pain and the ability to do activities of daily living. The low back score of Greenough has 13 parameters which assess patient in terms of pain, activities of daily living like sleeping, walking, sitting, travelling and dressing. Out of the total scores of 75 points, score >65 is excellent while <29 is said to be poor. The Frankel grading is classified from A to E depending upon the sensory and motor function with E as normal sensory and motor function and grade A as complete loss of sensory and motor function.

Statistical analysis was conducted with the statistical package for the social science system version SPSS 22.0. Continuous variables were presented as mean±Standard Deviation (SD) and categorical variables were presented as absolute numbers and percentage. Ordinal non-parametric data (Functional outcome) between the groups were compared using Mann-Whitney’s U-test. The pre and postoperative values of the scoring system were calculated using the unpaired t-test. The p-value of <0.05 was taken as statistically significant.

General anaesthesia and prone position were used in all the patients. All the surgeries were performed by single surgeon AB. Three doses of second generation cephalosporins (Inj Cefuroxime 1.5 g) were injected 30 minutes prior to the incision followed by two doses at 12 hourly intervals. A level proximal and distal to the fractured vertebrae was marked preoperatively with the help of C arm image intensifier in both the orthogonal views. After subcutaneous dissection, the transverse process and pedicles were cleared off the attachments. The entry at the pedicles was marked and four screws were inserted (two proximal and two distal to the fracture site). The decision regarding the need for decompression was made depending on the neurological status. Posterolateral fusion was done at this stage in Group B patients after harvesting the graft from the posterior iliac crest. The final reduction was done using the distractive forces at the longitudinal rods which were connected to the pedicle screws and the reduction was checked in C arm intraoperatively. Similar pain control management protocols were followed for all the patients. The patients were mobilised with the help of brace from postoperative day one. Neurological status along with radiological examination was done of all the patients at 3, 6 and 12 months respectively.

The mean age of patients in Group A was 36.81±4.21 years and 31.87±3.06 years in Group B. There were 17 (54.8%) males and 14 (45.2%) females in the present study. The most common mode of injury was RTA and L1/L2 vertebra was the most common vertebra to get fractured [Table/Fig-1].

There was a gradual improvement in Frankel scoring, reduction of the anterior vertebral height and kyphotic angle from preoperative status to subsequent follow-ups at one year [Table/Fig-2,3 and 4]. The Greenough low back outcome score was 45.25 in Group A and 46.10 in Group B cases which were not statistically significant (p=0.07). As per the modified Mcnab’s questionnaire 8 (50%) cases had excellent, 6 (37.5%) had good and 2 (12.55%) cases had poor functional outcome in Group A, while 9 (60%) cases had excellent, 4 (26.7%) had good and 2 (13.33%) cases had poor functional outcome in Group B postoperatively. One (6.25%) in Group A and 2 (13.33%) cases in Group B had superficial infection. All the three patients were treated with oral antibiotics. Screw loosening was apparent in 2 (12.5%) cases in Group A and 1 (6.66%) case in Group B respectively.

The thoracolumbar junction being the transition zone is more prone to injuries which can lead to deformity and even neurological deficits. The most commonly involved vertebra in Group A were L1 and L2 while T12 and L1 vertebrae were seen more frequently in patients of Group B. The findings of the present study were similar to others in literature [10,12,22]. There was a male predominance seen in the present study affecting 54.8% cases which were consistent with other studies [23,24]. Tezeren G et al., in their study of 42 patients compared the functional and radiological outcome between long-segment instrumentation with and without fusion, and found no significant difference [10]. We had no experience with long term segmental instrumentation in the present study. In a retrospective analysis by Hwang JU et al., comparing the results of short-segment fixation and fusion versus only fusion [12]. They concluded that fusion has better outcome in terms of kyphosis, pain and implant-related complications. Similarly, the rate of screw loosening was 12.5% in fixation group (12.5%) was more than the fixation and fusion group (6.66%) in the present study. In contrast, the other parameters like change in kyphosis angle and vertebral height had no significant difference.

Frankel scoring system helps to assess the severity of spinal injury with respect to the sensory and motor involvement. At the final follow-up, 87.5% cases in Group A and 86.6% cases in Group B had grade E Frankel’s grading. Thus, there was a gradual improvement in the sensory as well as motor status of the patients. The role of fusion has been controversial. Proponents of the non-fusion believe that it avoids the donor site morbidity, reduces the intraoperative surgical duration and blood loss while preserving the mobility of the adjacent segments [2]. These observations have been confirmed by Quian BP et al., Hwang JU et al., Yang X [7,12,23]. In the present study, the average surgical duration and intraoperative blood loss among Group A and Group B were statistically significant. The opponents, on the other hand, are of the view that the non-fused segments may alter the biomechanics caudal and cephalad to the fixed vertebra, it eliminates the low back pain symptoms and helps in achieving a better deformity correction [10,22,23].

The preoperative kyphotic angle in Group A was 27.69±6.720 which improved to 9.13±2.890 at the 12 month follow-up. Similarly, the preoperative kyphotic angle in Group B was 28.67±5.430 which reduced to 10.40±1.840 at the final follow-up. The change in the kyphotic angle was not found to be statistically significant (p=0.082). The low back outcome scale of Greenough and Fraser showed no difference between both the groups (p=0.07). In the present study, as per the modified Mcnab’s questionnaire, 87.09% of the cases had excellent to good results at the final follow-up. The posterolateral fusion group patients had no better outcome than the non-fusion group. A recent meta-analysis by Tian NF et al., showed no upper hand of fusion with fixation over fixation alone in unstable thoracolumbar burst fracture fixations. Similar findings were observed in the present study [25].

The present study has some limitations. Sample size determination was not done which remains the drawback of the present study. Also, the follow-up of the patients is short-term (one year) when the failures are not that prominent. The difference between the Greenough and Fraser score was not statistically significant which might be due to the low sample size which is a limitation of the present study.

The dilemma in treating the burst fractures of thoracolumbar spine still exists in terms of fusion versus non-fusion. The findings of the present study affirm that fusion is not superior to fixation for the thoracolumbar burst fractures. However, further meta-analysis and large randomised control trials are required for defining a treatment protocol.