Tuberculosis is a major healthcare problem in many developing countries, and tuberculosis of the spine (TBS) is one of the most common extra-pulmonary forms. More than two million people suffer from TBS worldwide [1], but with advancements in drug therapy, most cases can be dealt with ambulatory chemotherapy. Non-operatively treated patients have an average of 15° increase in kyphosis and if left untreated, 3%–5% of these cases gradually progress to develop severe kyphosis [2]. Overlooked diagnoses, poor accessibility to healthcare, and socio-economic factors are major reasons for the occurrence of such severe post-tubercular kyphosis (PTK).

PTK leads to significant physical and psychological morbidity. Persistent pain, costo-pelvic impingement, cardiopulmonary compromise, and late onset neurological deficits are problems associated with PTK. Poor cosmetic appearance associated with PTK is a major psychological burden due to a social stigma particularly in developing countries. Surgical correction of kyphosis will greatly improve these patients quality of life.

Untreated TBS, especially in children, leads to destruction of multiple vertebral bodies and subsequent shortening of the anterior column. Progressive kyphosis further adds to anterior column shortening and causes posterior column lengthening. Therefore, the goal of surgical procedures is to lengthen the anterior column, shorten the posterior column, or accomplish both. Surgical approaches available for treatment of kyphosis are: only anterior procedure [3] or only posterior procedures, like pedicle subtraction osteotomy (PSO) [4,5], or combined posterior-anterior procedures [6,7]. We describe our experience with single stage vertebral column resection (VCR) with sagittal rotation (SR) and controlled anterior opening and posterior closing (AOPC).

Between January 2007 and December 2014, we retrospectively reviewed the operating theatre reports, patient records, and radiographs of 21 consecutive patients who had surgery for PTK. We obtained approval from the institutional review committee of Orthopedics Seth GSMC and KEM Hospital (IRB approval no., 45/2015) before starting this study. Written and informed consent was given by patients for use of their clinical and radiological data for the purposes of this study.

Our indications for surgery were: presence of neurologic deficit (n=8), progressive kyphosis (n=9), intractable pain (n=3), and cardiopulmonary distress (n=1). We noted patient age, sex, diagnosis, levels involved, pre and postoperative angle of kyphosis (K angle), loss of correction at follow-up, pre- and postoperative neurological status (Frankel grading), pre- and postoperative pain assessed by Visual Analog Scale (VAS) score, estimated blood loss, and postoperative complications.

This study included 21 patients (eight males and 13 females) with a mean age of 21.90±7.85 years (range, 12–43 years). All patients had a minimum follow-up of 18 months. Average follow-up was 30.4 months (range, 18–96 months). No patient had osteoporosis (the T score of all patients was above −2.5 on the DEXA scan). Clinical and radiological data of patients is summarized in Table 1.

All patients had a history of childhood TBS and were treated with chemotherapy and mobilization with brace. The mean duration of chemotherapy was 9.3±3.12 months (range, 3–18 months). The resected material in one patient showed histopathological evidence of tuberculosis and subsequent culture showed multi-drug resistant tuberculosis. The patient was successfully treated with combination chemotherapy for 2 years per our institutional protocol. There was no clinical and radiological evidence of recurrent tuberculosis in any patient. The average number of vertebral bodies destroyed was 2.57±0.51 (range, 2–3).

The estimated blood loss was 917±144 mL (range, 700–1,300 mL) and average operative time was 444±59 minutes (range, 36–540 minutes). Anterior reconstruction was performed using tricortical iliac crest bone graft in four patients and cage in 17 patients. At the presentation to our clinic, mean preoperative kyphosis angle was 68.42°±13.23° (range, 40°–101°). Immediately postoperative mean K angle was 8.71°±4.58° (range, 2°–22°). Improvement in K angle postoperatively was statistically significant (p<0.001). The average percentage of correction was 87.11%±6.02% (range, 75%–96%). We achieved a mean correction of 59.71°±12.98° (range, 32°–97°). There was no significant difference in improvement of K angle depending on the number of vertebral bodies destroyed (p=0.16).

At the latest follow-up visit, the mean K angle was 12.28°±5.08° (range, 4°–26°) with an average loss of correction of 3.57°±2.03° (range, 1°–11°). The average correction at the latest follow-up visit was 81.68%±7.09% (range, 70%–94%). Part of this apparent loss in correction could be attributed to the fact that immediate postoperative X-rays were taken in the supine position while the X-rays at follow-up were taken in the standing position.

Thirteen patients did not have any preoperative neurological deficit, which did not change postoperatively. Out of five patients with Frankel grade C, four recovered completely (grade E) but one patient remained as grade C. The average time for recovery in these four patients was 17 weeks (range, 8–36 weeks). Out of the three patients with Frankel grade A, one patient recovered to grade D and was independently ambulatory with the aid of a walker in 6 months. The other two patients recovered to grade C (in 4 and 6 months, respectively). No patient had a deteriorated status following surgery.

Pain as judged by VAS score significantly improved from an average preoperative score of 6.38±0.92 (range, 5–8) to 1.38±0.49 (range, 1–2) at the final follow-up visit (p<0.01). There was no significant difference in improvement of VAS score depending on number of vertebral bodies destroyed (p=0.16).

Successful fusion was achieved in all cases. This was confirmed radiologically by the presence of bridging bone between the proximal and distal extent of the VCR (Fig. 1B, C).

Two patients experienced superficial wound maceration, which healed with daily cleaning and dressing. One patient developed persistent hypotension post operatively. He was managed in the intensive care unit with intravenous fluids, whole blood, and temporary inotropes along with strict monitoring. The patient eventually made an uneventful recovery. One patient developed a hemothorax postoperatively and was managed by chest drainage tube insertion and regular monitoring. The patient recovered without the need for any additional surgery. As a result of nerve root ligation, patients experienced hypoesthesia in the dermatomal distribution if the ligated nerve root. However, none of the patients complained of radicular pain. The patients regained near complete sensation 6–12 months following surgery, likely due to the overlapping of adjacent dermatomes.

Before the development of the halo-pelvic apparatus, treatment of severe PTK was difficult and dangerous. In 1974, Yau et al. [8] described the correction of PTK with spinal osteotomy, halo-pelvic distraction, followed by anterior and posterior fusion. The process was long and cumbersome and the average correction achieved was only 28.3% with serious complications including death of patients. However, with the advancement in surgical techniques and development of pedicle screws, new frontiers have opened up for the correction of PTK.

As described earlier, PTK leads to the shortening of anterior column and lengthening of the posterior column. Only anterior surgery offers advantage of direct under vision anterior debridement; however, achieving correction of severe kyphosis is difficult without strong posterior pedicular screw instrumentation. Moreover, due to old tuberculosis, there may be adhesion within the chest cavity making surgery difficult and is prone to complications. Only posterior osteotomy, like PSO, which shortens the posterior column, is also effective for correction of kyphosis. However, in PTK, due to significant anterior body loss with associated short anterior column, only posterior column shortening will lead to kinking of the cord and increases the risk of neurological deficit.

Osteotomy has the advantage of simultaneously lengthening the anterior column and shortening the posterior column, thus preventing excessive kinking of the spinal cord. One very important aspect of anterior lengthening in parallel to the posterior shortening is that this combination of maneuvers moves the fulcrum from the anterior edge of the vertebral column in a posterior direction. Thus, there is less shortening, less AP translation, and less angular realignment at the level of the spinal cord. This helps to prevent inadvertent spinal cord injury.

There is no doubt that neuromonitoring is the standard of care for severe deformity correction and is vitally important. Since we do not have a neuromonitoring facility in our hospital, we still rely on the age-old wake-up test. The process of instrumentation and VCR is essentially performed by an extremely experienced team of trained anesthesiologists and surgeons. After the temporary rod fixation, over screw VCR is performed diligently with due care. Just prior to the removal of the cortical shell of the internal gibbus, the feeling of the neural tissue on the anterior and anterolateral part of the cord is much like a taut band when it is palpated by a smooth penfield or a ball tip probe. The posterior dura feels soft and relaxed. Once the corrective maneuver is performed the anterior and anterolateral dura appears to be more lax compared to the previous pre-correction tautness. This is essentially because the corrective maneuver is a column shortening procedure. Rather, we should avoid the overcrowding of neural tissue, which usually happens during shortening procedures. This is apparent due to the appearance of wrinkles and insinuation. So, the surgeon can get some idea about the amount of stretch the neural tissue might experience during the described protocol to correct for the deformity. Since the wake-up test is performed after the completion of the deformity correction in absence of neuromonitoring, the only possible way to assess the hazardous stretch of the neural tissue is literally in the hands of the surgeon in a resource-poor setting. Anesthetists always maintain a systolic blood pressure of >90 mm Hg. Only spinal surgeons with sufficient experience and methodical training are permitted to perform this surgery.

Incomplete correction is defined as a correction of the deformity, in which the apex of the deformity has not been fully corrected. In severe kyphosis, the end plates of the vertebra proximal and distal to the excised wedge may not become parallel. Due to this, the spacer device used for the anterior reconstruction (cage or bone graft) will not have complete contact with the bone (Fig. 9B). If excessive compression is given, part of the cage might sink into vertebral body (Fig. 9C). Conversely, the desired correction is when the end plates of vertebra proximal and distal to the excised wedge are nearly parallel to each other (Fig. 9A). In cases of desired correction, the anterior spacer will have complete contact with the bone, thus improving the stability of the correction, and also allowing better surface area for vertebral union.

Bezer et al. [2] described transpedicular decancellation osteotomy for PTK and achieved a mean kyphosis correction of 17.1°. Kalra et al. [4] achieved a mean correction of 44.2° using PSO in PTK. Hong-Qi et al. [9] described a modified PSO and obtained a mean correction of 40.6°. Suk et al. [10] described their of posterior vertebral column resection (PVCR) used for correction in PTK, and although they were able to achieve a mean correction of 56°, mean blood loss was 2,980 mL and complications were reported in 10 out of their 25 patients. Lu et al. [6] described a technique of PVCR combined with manual intraoperative traction for correction of PTK. Two unscrubbed assistants applied longitudinal traction by pulling on the patient’s shoulders and ankles under the drapes provided manual traction. Although they achieved a good radiological correction (a mean correction of 80.1%), the average blood loss in their series was 2,604 mL, compared to 914 mL in our series and 820 mL and 782 mL in two other series of AOPC osteotomy. In our opinion, the control of manual traction used to correct deformity is difficult intraoperatively; inadvertent over distraction is possible, which can cause serious damage to spinal cord. We achieved a mean correction of 59.71° (87.10%) and encountered only two major complications, none of which required a secondary revision surgery. Improvement in pain as judged by the VAS score was similar to the series of cases discussed above. A brief comparison of results of our series in the context of the above series is shown in Table 2.

Kawahara et al. [5] was the first to describe AOPC osteotomy for non-infectious kyphosis in 2001. Rajasekaran et al. [11] first described a first series of AOPC osteotomies for PTK in 2010, achieving an average correction of 56.8%±14.6%. In 2014, Pang et al. [12] reported a series of 12 patients using same procedure and achieved a mean correction of 65.9%±6.35% compared to the 87.1%±6.02% in our series. We feel that this difference is primarily due to a difference in our respective techniques for anterior distraction. After excision of the wedge, Rajasekaran [7] and Pang et al. [12] used gradual posterior compression to cause simultaneous shortening of the posterior column and opening of the anterior column. In addition to posterior compression, we also used an interbody spreader anteriorly to further open the anterior column, thus achieving greater kyphosis correction. Chen et al. [3] described a similar procedure in 2011, demonstrating 72.17% correction in patients with severe angular kyphosis.

The surgical technique described in this paper can potentially be applied to any sharp angular kyphosis. The presence of osteoporosis may be considered a relative contraindication to this procedure as this surgery relies heavily on the good hold of the pedicle screws. Patients with osteoporosis may be treated appropriately prior to considering a deformity correction. Since none of our patients were above 50 years and majority of our patients (12 of 14) were less than 30 years old, we did not face this problem.

We must state that although this procedure is potentially less damaging to the spinal cord, serious complications may arise from the implementation of this procedure if performed by an inexperienced operator. Although we did not experience any major neurological complications, we would like to remind the readers that such corrective osteotomies have potential to cause serious neurological complications, including, but not limited to, irreversible paraplegia. We fully endorse the Scolio-1 Risk Trial and work by Lenke et al. [13] and always counsel the patient and their family regarding the substantial risk of neurological deficit, particularly as these deficits may not be just minor but could include paraplegia. Apart from counseling the patient and family about possible risks, we also actively involve them in vigorous preoperative chest physiotherapy and nutritional care, as most of these patients hail from poor socio-economic strata.

One limitation of our study is the small sample size. Also, due to absence of a control group, direct comparison with other procedures like PSO, modified PSO, or PVCR was not performed. Future prospective comparative studies may provide further insight into the advantages and potential fallacies of these procedures.

Simultaneous controlled anterior column lengthening and posterior column shortening by an AOPC osteotomy is an effective method for surgical correction of severe PTK. This procedure can be safely performed in a single stage and has been shown to have good clinical and radiological outcomes.