Management of Andersson lesions of spine: A systematic review of the existing literature

Abstract

Introduction

Andersson lesions also termed as aseptic spondylodiscitis, spinal pseudoarthrosis are known to occur in patients with ankylosing spondylitis. Trauma as well as inflammation has been cited as factors responsible for the causation of these lesions. A variety of surgical approaches have been described in the literature such as anterior, posterior, combined anterior and posterior, with or without reconstruction of the anterior column defect. Controversy still exists regarding the optimal management these lesions.

Objective

To address the optimal method of management, levels of instrumentation, requirement of fusion and anterior instrumentation and general epidemiological profile of the patients with Andersson lesions.

Materials and methods

An electronic search for studies on the surgical management of Andersson lesions of spine was performed. Quality assessment of the included articles was done by two independent authors according to the criteria used by researchers previously in systematic reviews.

Results

Males were found to have an increased incidence with the thoracolumbar junction being the most common level. Posterior approach was the most favoured with reconstruction of the gap in the anterior column. Posterior osteotomy with correction of deformity was done commonly for an optimal healing environment. Instrumenting 2–3 levels above and below the lesion is favoured by most.

Conclusion

Conservative management for Andersson lesions can be employed in the setting of acute trauma and stable fractures involving a single column. Surgical management of these lesions with a posterior long segment fixation and anterior column reconstruction is the favoured treatment in majority of the cases.

1. Introduction

Andersson lesion or aseptic spondylodiscitis is a destructive vertebral or disco-vertebral lesion occurring in the late stages of ankylosing spondylitis caused by trivial trauma.¹ It is either an erosive or sclerotic process extending either into the disc or adjacent bone. The incidence of Andersson lesion in patients with ankylosing spondylitis has been reported in the literature from 5% to 28% with male preponderance.²

Most of these lesions are seen at the junctional regions; thoracolumbar junction being most commonly involved region with one or more levels involved. The lesion was first described by Andersson in 1937.³ Cawley et al. classified these lesions into 3 groups according to the involvement of the discal surface of the vertebral rim (types A and B), the cartilaginous part of the vertebral end plates (type C and D) or both (type E).⁴

Many different etiologies have been proposed including inflammation, trauma, infections. Despite various possible etiologies, surgical management is indicated in lesions involving all the three columns of spine, associated painful deformities, neurological deficits. A multitude of management options have been described in the literature including anterior instrumentation with fusion, circumferential fusion, posterior short segment fixation, posterior long segment fixation with anterior instrumentation; however, a single consensus regarding the levels to be instrumented and the procedure of choice still could not be achieved. Similarly only fixation will suffice or whether debridement also is needed is also unclear.⁵

The authors of this article have employed a systemic review methodology to find out the various treatment options available for Andersson lesions in patients with ankylosing spondylitis and to reach a consensus regarding the optimal treatment method for these lesions.

2. Materials and methods

2.1. Inclusion/exclusion criteria

The inclusion criteria were defined under three comprehensive categories: study population and their demographic characters, independent variables and outcome measured. The study population included patients with ankylosing spondylitis having Andersson lesion in spine, demographic characters include gender and age of the patients. The independent variables included were the etiology of lesion, level of spine involved, management, levels included in instrumentation, duration of follow-up and complications. The outcome variables common to each study included in the review are solid bony fusion at the site of lesion, improvement in the VAS, ODI and neurological deficits.

2.2. Exclusion criteria

Case reports, technical reports as well as studies not mentioning the surgical management and outcomes were excluded from the review.

2.3. Search strategy and information sources

An electronic search of Pubmed, Embase and Cochrane data base was done up to March 2021. Both MeSH and Non-MeSH search terms used were ((Pseudoarthrosis) AND (surgery), ((Spondylitis, Ankylosing) AND (surgery), ((spine pseudoarthrosis) AND (ankylosing spondylitis)) AND (surgery), (Andersson lesion) AND (management), ((Andersson lesion) AND (ankylosing spondylitis)) AND (treatment), (((Andersson lesion) AND (spine)) AND (surgical management)) AND (fusion), ((spinal pseudoarthrosis) AND (ankylosing spondylitis)) AND (surgery). The references of the selected articles were also searched for additional studies which matched the inclusion criteria.

2.4. Data extraction

Data from each of the included studies was extracted independently by two reviewers as per the standard data extraction form for systematic reviews under the following headings (1) Author name, (2) Year of publication, (3) Population, (4) Intervention(s), (5) Comparisons, (6) Outcomes, (7) Results, (8) Authors main conclusions.

2.5. Data items

Articles describing the surgical management of Andersson lesion along with their functional and surgical outcomes were given weightage among the included articles. Articles in any language other than English and duplicated articles were excluded from the review. This was followed by abstract review where all the unrelated studies were excluded. At the final step, a detailed review of all the included studies was done and those in which surgical management of Andersson lesion or spine pseudoarthrosis in patients with ankylosing spondylitis and their outcomes was not mentioned were excluded. Any discrepancies or confusion between the two authors were resolved by including a third reviewer.

2.6. Assessment of bias

The bias among the included studies was assessed with the help of Risk of Bias In Non-randomized Studies (ROBINS-I) assessment tool suggested by Cochrane group.⁶ The above mentioned tool assess bias under the following domains: “Bias due to confounding”, “Bias in selection of participants in the study”, “Bias in classification of interventions”, “Bias due to deviations from intended interventions”, “Bias due to missing data”, “Bias in measurement of outcomes”, “Bias in selection of reported result”. The overall bias in the end is classified as Low/Moderate/Serious/Critical/NI (No Information). The assessment was done by two authors independently and a common consensus was reached following which the studies were included for the review. Any discrepancies between the author were, a third independent author was consulted. Risk of bias assessment was not blinded to the journal, institution or the authors.

2.7. Quality assessment

The studies included in the review were subject to quality assessment by two independent reviewers after blinding the authors and their affiliations. The methodology used for quality assessment of the included studies has been used beforehand by researchers in the quality assessment of systematic reviews.⁷ The studies were graded as “good” if the total score was more than 50% and “average” for a score less than 50%. Any discrepancies between the two authors were resolved by the third author.

2.8. Assessment of heterogeneity

A statistically significant heterogeneity in a study was indicated with a p-value of the χ² test less than 0.05. Clinical heterogeneity was assessed for all the retrieved studies. Pooling of the data was planned only if the data was appropriately homogenous.

3. Results

The PRISMA flowchart of study selection is shown in Fig. 1. A total of 73 studies were extracted an additional 7 studies were selected after reviewing the reference list of the selected articles. After removal of overlapping articles, irrelevant and duplicated studies a total of 15 articles were selected for the systematic review. All the included studies in the review were graded as “good” by the reviewers; however, none of the studies were prospective studies or RCTs and how the missing cases or lost to follow up was dealt with wasn't mentioned. The useful point among all the included studies were a clearly defined outcome variable with appropriate statistical analysis and conclusion addressing the outcome variable in question.

Fig. 1.

PRISMA flowchart for study section (26).

Risk of bias in the included studies: Overall, the risk of bias was high among the included studies.

Allocation bias: Since all the included studies were case series without any high quality evidence, the risk of allocation bias was high.

Blinding (Performance and detection bias): None of the included studies could be blinded for the patients and providers because of the kind of intervention provided. The outcome assessor was not blinded in either of the studies.

Incomplete outcome bias (Attrition bias): The attrition rate among the included studies was very less, causing the attrition bias to be less.

Selective reporting (Reporting bias): All the included studies seemed to be free of reporting bias.

3.1. Epidemiology

The mean age group of the patient population included among all the included studies was 48.12 ± 7.32 (range:35.7–61.1). It was seen that the prevalence of Andersson lesion in males was more compared to females for the same age group (Table 1).

Table 1.

Demographic profile of patients with Andersson lesion managed surgically.

Study/Year	Patients/Lesions	Mean age	Gender	Level of spine involved	Etiology
Fang et al. (1988) (10)	35 patients (40 lesions)	–	Males (31) Females (4)	Thoracolumbar (36) Lumbar (4)	Traumatic (7) Insidious (33)
Olerud et al. (1996) (6)	31 patients	60 years	Males (25) Females (6)	Cervical (19) Thoracolumbar (12)	Traumatic (26) Non-traumatic (5)
Chen et al. (2005) (19)	8 patients	49 years	Males (7) Females (1)	Thoracolumbar (7) Dorsal (1)	–
Chang et al. (2006) (14)	30 patients	41.7 years	Males (26) Females (4)	Thoracolumbar	–
Kim et al. (2007) (15)	12 patients (19 lesions)	42.8 years	Males (11) Females (1)	Thoracolumbar (10) Thoracic (1) Lumbar (7) Lumbosacral (1)	–
Dave et al. (2011) (16)	14 patients (15 lesions)	61.13 ± 19.74 years	Males (13) Females (1)	Thoracolumbar	Traumatic (5) Insidious (10)
Qian et al. (2012) (18)	7 patients	41.7 years	Males (6) Females (1)	Thoracolumbar	Traumatic (4) Insidious (3)
Zhang et al. (2015) (17)	17 patients	48.7 years	Males (12) Females (5)	Thoracolumbar	Traumatic (9) Insidious (3)
Zhang et al. (2016) (21)	13 patients	36.8 years	Males (9) Females (3)	Thoracolumbar	–
Liang et al. (2017) (20)	14 patients	35.7 ± 6.1 years	Males (13) Females (1)	Thoracolumbar	–
Reddy et al. (2017) (5)	20 patients	50.1 years	Males (18) Females (2)	Thoracolumbar (12) Lower thoracic (5) Lumbar (3)	–
Yang et al. (2017) (22)	12 patients	52.2 years	Males (10) Females (2)	Thoracolumbar	Traumatic (9) Insidious (3)
Ling et al. (2017) (23)	11 patients	53.1 years	Males (10) Females (1)	Thoracolumbar (7) Thoracic (2) Lumbar (2)	Traumatic (5) Insidious (6)
Wang et al. (2017) (24)	12 patients	49.6 years	Males (9) Females (3)	Thoracolumbar	Traumatic (3) Insidious (9)
Shaik et al. (2018) (13)	18 patients	54.6 years	Males (17) Females (1)	Thoracolumbar (12) Thoracic (3) Lumbar (3)	–
Wei et al. (2020)(25)	23 patients	44.8 years	Males (18) Females (5)	Thoracolumbar	–

Almost all the studies in the literature have cited thoracolumbar region as the most common site for Andersson lesions complicating ankylosing spondylitis except one study by Olerud et al.,⁸ patients with cervical spine fracture outnumbered those with thoracolumbar pathological fractures. Majority of the pseudoarthrosis lesions in the patients had trivial trauma such as fall from standing height as an inciting factor for the development of the lesion which again points out towards other factors such as osteoporosis or poor bone mineral density to be responsible.⁹ Fractures occurring due to trauma are less common in the cervical spine owing to less mechanical loading compared to the dorso-lumbar spine.

However, patients without a history of trauma also had these lesions in the spine so it could not be concretely concluded that a history of trauma has to be present for these lesions to develop.

Type of fracture in relation to the level of spine involved is shown in Table 2¹⁰

Table 2.

(10): Type of fracture and its location in the spine in Andersson lesion.

	Cervical	Thoracic	Lumbar	Sacral	Total
Extension	68 (68%)	15 (93.7%)	12 (100%)	1 (100%)	96 (74.4%)
Flexion	20 (20%)	0	0	0	20 (15.5%)
Compression	5 (5%)	1 (6.3%)	0	0	6 (4.7%)
Rotation	7 (7%)	0	0	0	7 (5.4%)
Total	100	16	12	1	129 (100%)

3.2. Conservative management

Conservative management is often considered as the first step in the treatment of these lesions. The optimal duration of therapy is not clear as almost no studies point out to a defined time limit, similarly conservative management has been described in the literature to be effective for upper dorsal spine lesions with an intact rib-sternum complex often described as the “fourth-column”.¹¹

Mobile segments of spine such as cervical and dorso-lumbar levels conservative management is less efficient. Despite normal healing capacity of bone in ankylosing spondylitis, healing of extensive lesions is less likely even after stabilizing with brace or plaster cast as minimal persistent motion will hinder the union at the site of lesion.⁴,¹² Nevertheless, successful healing of acute fractures in ankylosing spondylitis has been demonstrated with either halo or plaster immobilization.¹³

Pharmacological management with non-steroidal anti-inflammatory drugs is the mainstay of management during the active phase of disease in ankylosing spondylitis, however exercises as advised for patients with ankylosing spondylitis are contraindicated in patients with Andersson lesion complicating ankylosing spondylitis¹⁴

3.3. Anterior approach

In one of the studies by Fang et al.,¹² anterior spinal fusion was performed with good outcomes. Some surgeons advocated a combined anterior and posterior approach in the management of pseudoarthrosis of spine but this two-step procedure increases the overall operating time as well as increases the intra-op blood loss leading to a higher probability for surgical complications.

3.4. Posterior approach

With time, the trend of operative approach of choice was slowly getting shifted towards a single posterior approach⁵,¹⁵,¹⁶,¹⁷,¹⁸,¹⁹ to deal with Andersson lesion as it has its distinct advantages over anterior approach. In ankylosing spondylitis as the entire spine is fused with motion remaining at the pseudoarthrosis site with a long lever arm on both sides requiring long segment instrumentation, posterior approach is more often favoured because of surgeons familiarity with this approach as well as the co-morbidities of anterior approach can be avoided (Fig. 2).

Fig. 2.

40 yr male known case of Ankylosing Spondylitis presented with neck pain. X ray(a) and CT scan(b) shows Andersson lession at C4-5 level. Patient underwent posterior instrumented stabilization along with posterior/posterolateral fusion; Post-op x-ray lateral (c) and AP view(d).

3.5. Requirement of fusion

Most of the authors in their studies favoured anterior fusion at the pseudoarthrosis site be it either with a tricortical graft or a cage via a posterior approach with few studies going for a staged procedure.

Qian et al.,²⁰ in their study of 7 patients with Andersson lesion at the thoracolumbar junction did a staged approach. In the first stage a Pedicle Subtraction Osteotomy (PSO) was done at the pseudoarthrosis site followed by a second stage anterior surgery with a tricortical iliac crest at the site of lesion after debridement. Sound radiographic union was seen in all patients without any loss of kyphosis correction or requirement of second surgery.

Chen et al.,²¹ similarly went for first stage anterior tricortical graft fusion at the pseudoarthrosis site followed subsequently by a posterior instrumented stabilization. Good bony fusion was obtained in all the patients except one who had bone graft dislodgement, the patient was managed conservatively as he had no other symptoms. They concluded that anterior bone graft is necessary as it decreases stress on the internal fixation and provides cortical contact to buttress against the flexion moment on the instrumentation.

Kim et al.,¹⁷ in their retrospective study evaluating 19 lesions in 12 patients with sagittal plane deformity along with Andersson lesion went for a staged procedure in patients who had a severe kyphotic deformity. Smith-Peterson Osteotomy (SPO) with anterior interbody fusion was performed at the pseudoarthrosis site, patients who required correction of lumbar hypolordosis PSO was performed below L1 in addition to SPO at the pseudoarthrosis site. For this subset of patients anterior interbody fusion was performed after a period of time as a staged intervention. All patients had a sound union at the pseudoarthrosis site.

The capability of ankylosing spondylitis to unite despite not augmenting the fixation with an anterior bone graft has been exploited by some authors. Shaik et al.,¹⁵ in their retrospective study of 18 patients did a posterior long segment fusion without anterior bone grafting or fusion with a cage. They found out that anterior column defects as large as 2.5 cms filled up as early as 8 months post fusion and completely fused by 2 years post-surgery with good fusion at the pseudoarthrosis site; however, whether any deformity was present at the site of pseudoarthrosis was not reported in their study.

Similarly, Liang et al.,²² in their retrospective study of 14 patients with Andersson lesion with severe kyphosis did a posterior wedge osteotomy without anterior cage or bone graft fusion. There was a statistically significant correction of kyphotic deformity at the involved levels with Grade 1 fusion achieved at the pseudoarthrosis site according to the criteria described by Bridwell et al. The authors concluded that patients with a severe kyphotic deformity with Andersson lesion can benefit from a posterior wedge osteotomy and debridement of the lesion.

Chang et al.,¹⁶ in their retrospective review of 30 patients with Andersson lesion did a posterior corrective osteotomy without anterior spine fusion. A substantial amount of correction was achieved by them with bony fusion in all their patients with no loss of correction on follow-up but the grading system which the authors used for assessing bone fusion was not explicit.

3.6. Requirement of an osteotomy

Almost all of the patients with ankylosing spondylitis present with some amount of deformity owing to the fused spine. Andersson lesion or pseudoarthrosis at a particular level tends to accentuate the deformity requiring its correction in the form of an osteotomy through the lesion (Fig. 3).

Fig. 3.

45 year female known case of Ankylosing Spondylitis with positive neck sagittal balance. Pre-op x ray AP (a), Lateral (b) CT Scan (c) suggestive of Andersson lession at D12-L1(yellow arrow). Patient underwent deformity correction using Bone-disc-Bone osteotomy through the Andersson lession (d). One year follow-up image (e) showing good fusion at osteotomy site.

Closing or opening wedge osteotomy at the site of pseudoarthrosis has been described in many studies with good outcomes with respect to deformity correction as well as achieving fusion at the site of pseudoarthrosis. There are only a few studies where the authors instead went for a debridement of the pseudoarthrosis site with anterior bone graft or cage placement to achieve fusion.

The results of the studies are shown in Table 3.

Table 3.

Surgical management of Andersson lesion as described in the included studies.

Sl. no.	Study/Year	Operative approach	Fusion	Requirement of an osteotomy	Levels of fixation
1.	Fang et al. (1988) (10)	Anterior	Iliac crest bone grafting	–	–
2.	Olerud et al. (1996) (6)	Both anterior and posterior approach for cervical spine Posterior approach for thoracolumbar spine	–	-	-
3.	Chen et al. (2005) (19)	Combined anterior and posterior	Anterior tricortical iliac crest graft	For kyphotic angles more than 45°	2 levels up and 2 levels below the level of lesion
4.	Chang et al. (2006) (14)	Posterior	No anterior fusion done	Open wedge osteotomy. Mean pre-op kyphosis 45.5°	-
5.	Kim et al. (2007) (15)	Posterior	With either bone graft or titanium cages	SPO for thoracic kyphosis more than 45°. SPO + PSO for combined thoracic hyperkyphosis and lumbar hypolordosis	-
6.	Dave et al. (2011) (16)	Posterior	Transpedicular bone grafting	-	1 level above and below the lesion
7.	Qian et al. (2012) (18)	Combined anterior and posterior	Iliac crest bone grafting	PSO at the level of deformity	-
8.	Zhang et al. (2015) (17)	Posterior	Bone graft and cage	Pedicle subtraction and disc resection osteotomy	2-3 levels above and below the lesion
9.	Zhang et al. (2016) (21)	Posterior	Bone graft for closure of osteotomy site	Osteotomy through the fracture gap	2-3 levels above and below the lesion
10.	Liang et al. (2017) (20)	Posterior	-	Posterior wedge osteotomy	-
11.	Reddy et al. (2017) (5)	Posterior	Cage with bone graft in the anterior defect	-	6.2 ± 2.4
12.	Yang et al. (2017) (22)	Posterior	PEEK cage for interbody fusion	Posterior transpseudoarthrosis oblique osteotomy	3 levels above and below the lesion
13.	Ling et al. (2017) (23)	Posterior	Bone graft for anterior defect	Grade 4 osteotomy (Pedicle/partial body and disc)	-
14.	Wang et al. (2017) (24)	Posterior	Bone graft for anterior defect	Posterior transpseudoarthrosis osteotomy	-
15.	Shaik et al. (2018) (13)	Posterior	-	-	2-3 levels above and below the lesion
16.	Wei et al. (2020) (25)	Posterior	Bone graft for the anterior defect	Modified posterior wedge osteotomy	2-3 levels above and below the lesion

3.7. Levels of fixation

No fixed criteria exist regarding short or long segment fixation and the number of levels that should be included above the lesion. Traditionally as the lever arm of the spine is long in ankylosing spondylitis and the only site with motion is the site of pseudoarthrosis, long segment fixations are always favoured; however, these have their own drawbacks in terms of operative time, blood loss and complications.

Dave et al.,¹⁸ did a short segment posterior fixation with anterior bone grafting with good results in terms of fusion across the pseudoarthrosis site without performing any correction of the kyphotic deformity.

In the studies included here in this review, majority of the authors have opted for long segment fixation except a few but the data on the levels of fixation is very heterogenous to specifically conclude anything concretely. One subjective assumption, which holds true for most studies that extending the instrumentation 2–3 levels above and below the lesion should be adequate for providing a good healing environment to both the osteotomy if done as well as the fusion.

The surgical and functional outcomes of the studies included in the review are shown in Table 4, Table 5.

Table 4.

Surgical and Functional outcomes among the included studies.

Study	Approach	Blood loss (in ml)	Duration of surgery (in mins)	Complications	Post-operative outcomes
Wei et al. (2020)	Posterior	488.5 (205–880)	205.4 (115–375)	Dural tears in 3 patients	VAS scores: Pre-op: 7.52 ± 1.31 Post-op: 1.70 ± 0.70
					ASQoL scores: Pre-op: 13.87 ± 1.89 Post-op: 7.22 ± 1.24
Shaik et al. (2018)	Posterior	533.3 ± 238.6	192.7 ± 41.6	–	VAS scores: Pre-op: 8.5 Final follow-up: 1.6
					ODI scores: Pre-op: 72 Final follow-up: 17
Wang et	Posterior	469.2 ± 251.0	247.9 ± 60.7	–	VAS scores:
al.					Pre-op: 8.58 ± 1.08
(2017)					Final follow-up:
					1.75 ± 0.75
Ling et al. (2017)	Posterior	1009	268.6	Dural tears in 3 patients	VAS scores: Pre-op: 6.7 Final follow-up: 0.27
Yang et al. (2017)	Posterior	817.5	201.9	1 dural tear 1 screw pull-out	VAS scores: Pre-op: 6.7 Final follow-up: 1.1
Rajoli et al. (2017)	Posterior	452	112	Dural tear: 3 patients Pleural tear: 1 patient Post-op pneumoencephalus: 1 patient Post-op hypoaesthesia in sacrificed nerve root dermatome: 3 patients Incomplete neurodeficits: 2 patients (Immediate post-op) Superficial surgical site infection: 3 patients	VAS scores: Pre-op: 8.2 ± 0.5 Final follow-up: 2.4 ± 0.6 ASQoL scores: Pre-op: 14.3 ± 2.08 Final follow-up: 7.9 ± 1.48 ODI scores; Pre-op: 62.7 ± 7.9 Final follow-up: 18.5 ± 7.5
Liang et	Posterior	1066.1 ± 466.7	279.4 ± 32.9	Dural tear: 1 patient	VAS scores:
al.				Pneumonia: 1 patient	Pre-op: 6.7 ±
(2017)					0.8
					Final follow-up:
					0.75 ± 0.6
					ODI scores:
					Pre-op: 60.56 ± 15.1
					Final follow-up:
					23.46 ± 8.2
Zhang et	Posterior	1100	280	Wound infection: 1	VAS scores:
al. (2016)					Pre-op: 7.2 ± 1.2
					Final follow-up:
					2.1 ± 1.1
Zhang et	Posterior	876	219	Dural tear: 3 patients	VAS scores:
al. (2015)				Pneumonia: 1 patient	Pre-op: 6.4 ± 1.1
					Final follow-up:
					1.1 ± 1.0
					ODI scores: Pre-op: 50.9 ± 10.4 Final follow-up: 16.9 ± 8.3
Qian et al. (2012)	Posterior (PSO) + Anterior	2200 (PSO) 470 (Anterior)	250 (PSO) 210 (Anterior)	None	VAS scores: Pre-op: 7.1 Final follow-up: 0
Dave et al. (2011)	Posterior	–	–	Dural tears: 2 patients Implant extension: 4 patients Superficial skin necrosis: 1 patient	VAS scores: Pre-op: 9.13 ± 1.82 Post-op: 3.47 ± 1.26 ASQoL scores: Pre-op: 13.93 ± 7.06 Post-op: 5.33 ± 9.52
Kim et al. (2007)	Posterior Smith Peterson osteotomy Pedicle subtraction osteotomy for severe kyphotic deformities in 6 patients	–	–	Dural tears: 3 patients Post-op radiculopathy: 2 patients Early deep wound infection: 1 patient	VAS scores: Pre-op: 7.9 Final follow-up: 3.1
Chen et al. (2005)	Anterior debridement with interbody fusion followed by posterior stabilization	3500	288	–	VAS scores: Pre-op: 67.75 ± 8.0 Final follow-up: 23.50 ± 5.29
Olerud et al. (1996)	Anterior + Posterior for cervical spine Posterior only for thoracolumbar spine	3500	240	Loss of fixation: 6 patients Neurological deterioration: 4 patients Infection: 2 patients Screw fracture: 1 patient Implant malposition: 1 patient Post-op fever: 1 patient Per-op haemorrhage: 1 patient	–

Table 5.

Comparison of radiographic parameters between pre-op and final post-op follow-up.

Study	Thoracic kyphosis (TK)	Thoraco-lumbar kyphosis(TLK)	Lumbar lordosis (LL)	SVA	Global kyphosis (GK)	Pelvic incidence (PI)	Pelvic tilt (PT)	Sacral slope (SS)	Local kyphosis (LK)
Wei et al. (2020)	Pre-op: 52.30± 17.62° Final follow- up: 28.45 ± 6.63°	Pre-op: 40.03 ± 17.61° Final follow- up: 28.45 ± 6.63°	Pre-op: −29.56 ± 9.73° Final follow-up: −20.73 ± 10.27°	Pre-op: 11.82 ± 4.55 cm Final follow-up: 5.03 ± 2.29 cm	–	–	–	–	–
Ling et al. (2017)	Pre-op: 48.1° Final follow- up: 35.6°	–	Pre-op: 19.1° Final follow-up: 3.5°	Pre-op: 9.1 cm Final follow-up: 5.5 cm	Pre-op: 72.3° Final follow-up: 47.5°	Pre-op: 49.2° Final follow- up: 50.3°	Pre-op: 25.8° Final follow- up: 19.6°	–	–
Yang et al. (2017)	–	Pre-op: 30.3° Final follow- up: 8.0°	–	–	–	–	–	–	–
Rajoli et al. (2017)	–	–	–	–	–	–	–	–	Pre-op: 27° Final follow- up: 18.1°
Liang et al. (2017)	–	Pre-op: 44.4 ± 12° Final follow- up: 6.6 ± 13.5°	Pre-op: 51.7 ± 15.6° Final follow-up: 7.1 ± 19.5°	–	Pre-op: 60.6 ± 28.3° Final follow-up: 20.3 ± 10.3°	–	Pre-op: 38.1 ± 14.8° Final follow- up: 20.2 ± 8°	Pre-op: 10.8 ± 14° Final follow-up: 25.7 ± 9.7°	Pre-op: 51.7 ± 15.6° Final follow- up: 7.1 ± 19.5°
Zhang et al. (2016)	–	Pre-op: 51.0 ± 9.9° Final follow- up: 24 ± 8.4°	Pre-op: 5.7 ± 23.2° Final follow-up: 18.8 ± 21.6°	Pre-op: 166 ± 37 mm Final follow-up: 87 ± 29 mm	Pre-op: 55.8 ± 11.0° Final follow-up: 26.4 ± 9.4°	–	–	–	–
Zhang et al. (2015)	–	–	–	Pre-op: 153.7 mm Post-op: 41.1 mm	Mean GK: Pre-op: 59.1° Post-op: 24.7° Regional GK: Pre-op: 29.1° Post-op: 5.9°	–	–	–	–
Qian et al. (2012)	Pre-op: 56.1° Post-op: 48.1°	–	Pre-op: −10.1° Post-op: −42.2°	Pre-op: 103.5 mm Post-op: 10.4 mm	Pre-op: 75° Post-op: 30°	–	–	–	Pre-op: 45° Post-op: 8.8°
Kim at al. (2007)	Pre-op: 57.5° Final follow- up: 48.2°	–	Pre-op: −17.1° Final follow-up: −35.5°	Pre-op: 18.8 cm Final follow-up: 3.6 cm	–	–	–	–	–
Chang et al. (2006)	–	–	–	–	–	–	–	–	Pre-op: 45.5° Final follow- up: 9.5°
Chen et al. (2005)	–	–	–	–	–	–	–	–	Pre-op: 41.3° Final follow- up: 29.5°

4. Discussion

Andersson lesion of spine have been described in the literature since a long time but till date controversy exists regarding the optimal management of these lesions. Surgical management of these lesions have been shaded by a number of questions such as ‘what is the approach of choice’, ‘whether fusion should be done for the defect after correction of the deformity’, ‘how many levels should be included in the instrumentation’. These voids in the literature were further supplemented by the lack of adequate evidence and systematic reviews with only case reports being reported initially.^5,8,14, 15, 16, 17 Gradually researchers began studying the nature of these lesions, their demographic profile and with the evolution of spine instrumentation surgical management of these lesions also evolved.

In our review it was seen that gender wise distribution of these lesions, males were most often affected with these lesions compared to females. Historically ankylosing spondylitis has been reported as a male predominant disease and the findings in our review is in accordance with the results of all the studies included. The increased incidence of these lesions in males could be attributed to increased prevalence of ankylosing spondylitis in males.

The region of spine most frequently involved is the thoracolumbar region which was seen in most of the studies with a few studies reporting the lesion in thoracic and cervical spine. The plausible explanation for the increased prevalence of these lesions in the thoracolumbar junction could be because of the increased amount of stress these areas bear as compared to the cervical and the relatively rigid thoracic spine.⁹

Trauma, inflammation, osteoporosis as well as stress all have been implicated in the causation of these lesions, the exact causative factor is still an enigma. Majority of the studies in our review have reported trivial trauma to be the inciting factor but other factors such as poor bone mineral density, increased frequency of fall due to the deformities should also be considered. So, trauma as the major causative factor could not be definitively stressed upon in our study.

Repeated stress can also lead to fracture of ankylosed spine similar to long bones, part of spine proximal to the lumbosacral junction also suffer from increased stress apart from the thoracolumbar junction. Stress fractures may initially occur in the posterior elements even before the appearance of lesions in the anterior column as the thoracolumbar and lumbar spine disc spaces are exposed to increased shear forces under the effect of gravity in a kyphotic spine.¹²

Hyper-extension moment has been cited as the most important mechanism implicated in causing fractures in patients with ankylosed spine, other forces such as flexion, rotation and compression type injuries are also described in the literature.

Conservative management is sought as the first alternative in acute fractures of the spine as well stable fractures involving one column of spine with studies suggesting good healing^11,13 but in mobile and junctional areas of the spine where shear forces as well as persistent micromotion is still possible despite immobilization with brace or plaster cast conservative management is less efficient.⁴,¹⁰

Surgical management along with their outcomes are increasingly being reported in the literature in the management of Andersson lesions. Different inferences have been provided by authors with regards to the ideal approach for surgical management of these lesions. To begin with a few studies⁸,¹², reported the results of anterior spine fusion in patients with Andersson lesions without instrumentation. They concluded that for cervical spine a combined anterior and posterior approach and for thoracolumbar spine a posterior approach with bone grafting of the anterior defect should be done for better results.

A slow and gradual trend towards an all-posterior approach for the treatment of these lesions was seen with good functional results and solid bony fusion across the pseudoarthrosis site.

Majority of the studies included in our review have supported the use of a bone graft for the defect left after correction of the deformity and instrumentation except a few studies where bone graft was not used for achieving fusion because of the inherent capability of ankylosing spondylitis to form bone, fusion easily occurs across the defects left after debridement of the lesion and correction of the deformity according to the authors. But the consensus of majority was that an anterior column support either with a cage or a bone graft with posterior instrumentation will provide a conducive environment for the healing of the lesion as the graft will be in compression with anterior column support preventing implant failure.⁵

Controversy still exists with the level of instrumentation, in terms of the number of segments of spine the instrument should span. Almost all the authors of the studies included in the review have opted for a long segment instrumentation to effectively counteract the movement at the pseudoarthrosis site except for one study¹⁸ where they have attempted a short segment instrumentation of the pseudoarthrosis lesion. They did not attempt to correct the deformity and no mention has been made regarding the anterior column defect after the instrumentation. The amount of post-op kyphosis as well as the anterior column defect are important predictors of the success of the instrumentation post-operative. Despite this being the limitation, they have reported good outcomes in their study without any major implant failure. The major consensus is still in the favour of a long segment instrumentation at least 2–3 levels above and below the lesion in order to give a biomechanically stable environment for the lesion to heal.

The current review highlights many contentious questions concerned with the surgical management of Andersson lesions. The void in the literature is primarily because of the lack of good quality studies as well as only case series and case reports being available on the topic. Further there is still a dearth in the literature with respect to the management of these lesions in cervical spine as all the studies available focusses primarily on the lesions involving the thoracolumbar spine. Ankylosing spondylitis is a disabling disease with an increased propensity of the patients to sustain fractures due to repetitive falls. Andersson lesion presenting with complaints of back pain, inability to stand, neurological deficits and progressive deformities require surgical management and identifying the optimal method of management of these lesions will provide the necessary assistance to a surgeon involved in the management of these cases.

5. Conclusion

A wide variation exists in the surgical management of Andersson lesions. Our review identified that incidence of these lesions is more in males compared to females with thoracolumbar region being the most commonly involved region. Trauma along with other factors such as poor bone mineral density and increased propensity to fall due to the deformity are the causative factors. Posterior approach with a long segment instrumentation including 2–3 levels above and below the lesion to be the most favoured approach among the surgeons. Reconstruction of the anterior column gap with bone graft or cage provides a biomechanically stable environment for the lesions to heal.

Funding

Each author certifies that he or she has no commercial associations that might pose a conflict of interest in connection with the submitted article.