The aim of this study is to compare the extent of bone destruction and radiologic parameters in human immunodeficiency virus (HIV)-negative and -positive patients with spinal tuberculosis (TB). The rate of HIV co-infection among patients with TB is rising in sub-Saharan Africa. There are conflicting views in the clinical literature regarding the pathology and the extent of bone destruction in HIV-positive patients with spinal TB. This study makes a contribution to the literature addressing this problem.
Methods
A prospective multicentre study was conducted on 61 consecutive patients who underwent spinal surgery for spinal TB. Clinical, laboratory and radiographic parameters, including MRI, were examined in both HIV-negative and HIV-positive patients.
Results
A total of 31 males and 30 females participated in the study. Seventy percent (43 patients) were HIV-positive and 18 were HIV-negative. There was no difference in clinical parameters between the two patient groups. The amount of vertebral body loss was significantly higher in HIV-positive patients (p = 0.01). The amount of pus was marginally higher in HIV-positive patients (p = 0.16). There was no statistically significant difference in the kyphosis, skip lesions and cord compression between the two groups.
Conclusion
Human immunodeficiency virus-positive patients exhibit more vertebral body destruction and a marginal increase in pus formation. This enhanced bone destruction is attributed to the activation of the receptor activator of the nuclear factor-κ ligand pathway by cytokines, a process that is more pronounced in the context of HIV infection.
Contribution
This clinical study supports the findings by laboratory studies that there is more bone destruction in HIV-positive patients than in HIV-negative patients with tuberculosis of the spine.
spinal tuberculosishuman immunodeficiency virusspine radiographsbone destructionpus formationFunding information The authors received no financial support for the research, authorship or publication of this article.Introduction
Tuberculosis (TB) remains a significant global health concern. In 2019, 10 million people developed TB and 1.4 million succumbed to the disease.1 Extrapulmonary TB is estimated to account for up to 20% of TB infections, 10% of these being musculoskeletal TB. The spine is affected in more than 50% of musculoskeletal TB extrapulmonary cases. Spinal TB accounts for 2% of all TB cases.2
The typical lesion of spinal TB on plain radiographs is a disease affecting two contiguous vertebral bodies with narrowing of the intervening disc space because of the collapse of the disc into the softened peridiscal vertebral bone. There are atypical lesions where the disease affects a single vertebral body or the posterior elements of the spine.3
Tuberculosis co-infection with human immunodeficiency virus (HIV) is now very common, occurring in more than 50% of patients with spinal TB in endemic areas.4
Clinical studies4,5,6 and studies from medical scientists7,8 provide contrasting views when comparing the extent of bone destruction and pus formation in HIV-negative and HIV-positive patients with spinal TB. Laboratory studies suggest more bone destruction in HIV-positive patients, and clinical studies report more destruction in HIV-negative patients.
The aim of this prospective clinical study is to compare the extent of bone destruction and radiologic features between HIV-negative and HIV-positive patients with spinal TB.
Research methods and design
A prospective, multicentre study of 61 consecutive adult patients aged 18 years or older undergoing surgical treatment for spinal TB between May 2019 and August 2020 was conducted. The indications for surgery were severe neurologic deficit, worsening neurology after 6 weeks of medical treatment and no response to treatment.9,10 The surgery was performed within 3 months of onset of symptoms in all patients. All the patients had been on TB treatment for at least 6 weeks. The TB diagnosis was made on a combination of a history of chronic back pain, neurologic deficit, radiographs and magnetic resonance imaging (MRI). These findings were consistent with TB and were later confirmed by histology, culture and the GeneXpert test. We excluded patients with sputum-positive pulmonary TB because of risk of postoperative pulmonary complications in this group. Also excluded were those with immunosuppressive conditions other than HIV – such as diabetes and rheumatoid arthritis – and individuals on chronic steroid therapy.
Ethical clearance was obtained from the university’s ethics committee (Ref no. No.73/2019). Informed consent was obtained from each patient enrolled in the study.
We recorded patient demographics, HIV status, Cluster of Differentiation 4 (CD4) count, viral load, Erythrocyte Sedimentation Rate (ESR), serum albumin levels, chest and spine radiographs and neurologic status using the American Spinal Cord Injury Association (ASIA) Scale.
On the radiographs, we recorded the anatomic area of the spine affected by the disease and the angle of the kyphosis using the Cobbs method.11 A radiologist examined the chest radiographs for pulmonary involvement.
An experienced musculoskeletal radiologist examined the MRI. The whole spine was assessed for skip lesions on the sagittal T1 weighted (T1W) sequences. T2 weighted (T2W) sequences and T1W fat-saturated (F/S) post-contrast sequences were used to better define the lesions, the involvement of posterior elements and the number of vertebrae involved. The volume of the abscess was measured by multiplying the following three values on the transverse images: the width and Anterio-posterior (AP) extent of the pus and the height of the lesion on the sagittal images.4,5 The loss of vertebral body height was assessed by dividing the vertical height of the vertebra on the lateral view into tenths based on the adjacent normal vertebra. The vertebral height loss was then estimated by measuring the lost tenths in each involved vertebra, expressed as a ratio of the normal vertebra.11
Statistical analysis was performed using STATA 17/18 software. Descriptive summaries were initially presented. This was followed by using the Student t-test for unequal variances to compare the results between the two groups for the continuous characteristics and an equivalent t-test for comparing proportions. A p-value of < 0.05 was considered to be significant.
Ethical considerations
Ethical clearance to conduct this study was obtained from the University of Pretoria, Faculty of Health Sciences Research Ethics Committee on 28 March 2019 (No. 73/2019).
Results
Out of the 61 patients enrolled for the study, 18 (30%) were HIV-negative and 43 (70%) were HIV-positive. Thirty-one of the patients were male and 30 were female. The age range was 25–64 years, with a mean of 42.8 years.
The demographic details of the two patient groups are presented in Table 1a and Table 1b. There is no statistically significant difference between the parameters that were measured.
Clinical parameters in human immunodeficiency virus-negative and human immunodeficiency virus-positive spinal tuberculosis patients.
Parameter
HIV-Negative
HIV-Positive
n
%
n
%
Patients
18
30
43
70
Mean age (years)
43.3
-
42.6
-
ASIA Scale
A
2
11.1
3
6.9
B
2
11.1
2
4.6
C
3
16.6
14
32.5
D
3
16.6
12
27.9
E
8
44.4
12
27.9
Pre-op TB Rx
11†
61
37‡
62
Viral load
-
-
31/43§
72
TB, tuberculosis; HIV, human immunodeficiency virus; AISA, American Spinal Cord Injury Association.
, for 2.6 months;
, for 3.7 months;
, undetected in 31/43 patients.
Clinical parameters in human immunodeficiency virus-negative and human immunodeficiency virus-positive spinal tuberculosis patients.
Parameter
Mean
s.e.
95% CI
Mean
s.e.
95% CI
p-value
Significance
ESR (mm/hr)
56.17
8.07
38.2–73.2
56.17
16.2
26.2–53.4
0.87
Non-sig
CRP (mg/L)
42.94
9.13
23.7–62.2
41.50
5.9
29.6–53.4
0.89
Non-sig
Albumin (g/L)
34.44
1.43
31.4–37.5
34.91
0.87
33.2–36.7
0.78
Non-sig
Platelets (109/L)
341.78
27.92
82.9–400.7
351.37
18.1
314.8–387.9
0.78
Non-sig
Hb (g/dL)
12.12
0.58
11.0–13.2
12.14
0.27
11.6–12.7
0.97
Non-sig
WCC (109/L)
6.79
0.52
5.8–8.0
5.89
0.41
5.1–6.7
0.21
Non-sig
CD4 (cells/µL)
-
-
-
385.00
38.0
308.1–461.9
n/a
-
s.e., standard error; ESR, erythrocyte sedimentation rate; CRP, C-Reactive protein; Hb, haemoglobin; WCC, white cell count.
On the chest radiographs, 46 patients (75.4%) had a clear chest. Thirteen (21.32%) had active disease, and two (3.28%) had evidence of fibrosis compatible with healed pulmonary TB.
GeneXpert test was positive in 85.45% of the patients. Histology revealed granulomas, Langhans’ giant cells and necrosis in various combinations in all patients. Culture was positive in 48.3% of patients after an average incubation period of 18 days. The Ziehl–Neelsen stain was positive in 15.5% of the patients.
Spinal radiographs demonstrated predominant involvement of the thoracic spine in 27 patients (44.2%) and the lumbar spine in 24 patients (39.3%). The rest of the patients showed involvement of thoraco-lumbar spine (6), cervical spine (3) and lumbosacral spine (1), as illustrated in Figure 1.
Distribution of tuberculosis in the various spinal areas.
The typical radiographic pattern of a two-body disease with narrowing of the intervening disc space, as illustrated in Figure 2, was observed in 45 patients (73.77%) and the atypical pattern in 16 patients (26.23%).
A plain lateral radiograph shows spinal tuberculosis involving the adjacent vertebral bodies of L3 and L4, with narrowing of the intervening disc space. The T1-weighted post-contrast magnetic resonance imaging (MRI) reveals ring enhancement consistent with an abscess. (b) A lateral radiograph demonstrates an atypical spinal tuberculosis (TB) lesion with symmetrical involvement of the vertebral body (indicated by the arrow). The post-contrast MRI image again shows ring enhancement, confirming TB.
Of the 18 HIV-negative patients, 14 (77.8%) had a typical radiographic pattern and four (22.2%) had an atypical pattern. In the HIV-positive group, 31 (72.0%) had a typical pattern and 12 (27.0%) had an atypical pattern. There was no statistically significant difference in the distribution of the radiographic pattern between the two groups (p = 0.64).
Skip lesions were assessed on a whole spine MRI. Four patients did not have a whole spine MRI. Skip lesions were found in six patients (10.53%). One patient in the 15 HIV-negative patients had a skip lesion (6.25%). Additionally, five skip lesions were found in the HIV-positive patients (11.9%). There is no statistical evidence of the association of skip lesions with the HIV status of patients (p = 0.66). Figure 3 depicts an example of a patient with clinical and radiologic thoracic spine lesions, where a whole spine MRI showed a skip lesion in the lumbar spine.
A patient presented with thoracic back pain and a thoracic gibbus. Antero-posterior (AP) and lateral radiographs of the thoracic spine demonstrate features suggestive of spinal tuberculosis (TB). The AP view (a) shows an abscess (indicated by the arrow), while the lateral view (b) reveals vertebral collapse. A whole spine T1W magnetic resonance imaging (MRI) sequence identifies skip lesions at T8 and L3 (c). Both lesions were confirmed as TB on T1W F/S post-contrast MRI sequences of the thoracic (d) and lumbar spine (e).
The results of the comparison of the radiographic parameters of the mean kyphotic angle, mean number of vertebrae involved, mean volume of pus and the mean loss of vertebral body height between the two patient groups are summarised in Table 2. There was significant vertebral body loss in HIV-positive patients. It is also worth noting that although there is almost double the amount of pus in HIV-positive patients, this is not of statistical significance. There was no statistical difference in the extent of spinal cord compression (p = 0.81).
Radiographic features of spinal tuberculosis in human immunodeficiency virus-negative and human immunodeficiency virus-positive patients.
Radiographic/MRI parameter
HIV-Negative
HIV-Positive
Statistical significance
Mean
s.e.
95% CI
Mean
s.e.
95% CI
p-value
Kyphosis (degrees)
25.380
2.90
19.26–31.50
25.090
2.370
20.29–29.89
0.32
Non-sig
No vertebrae involved
2.55
0.24
2.03–3.07
3.27
0.30
2.66–3.89
0.07
Non-sig
Volume of pus (mm3)
63.30
19.11
22.5–104
109.92
28.3
52.60–167
0.16
Non-sig
Vertebral body height (ratio)
0.968
0.10
0.73–1.1
1.33
0.10
1.2–1.5
0.01
Sig
MRI, magnetic resonance imaging; s.e., standard error.
Discussion
The World Health Organization reports that 8.2% of all TB cases occur in people living with HIV.1 In 2012, Anley et al. reported that 40% of their spinal TB patients were HIV-positive.4 In a 2018 retrospective study of 274 patients, 76% of patients with spinal TB were HIV-positive.6 In this series, 9 years after the report by Anley et al., 70.5% of the patients with spinal TB were HIV-positive.
Our results show no difference in age and gender distribution in HIV-negative and HIV-positive patients. This is a similar finding to the study by Marais et al.6 In another study, HIV-positive patients tended to be younger.3
In 79% of the patients, the disease affected the thoracic and lumbar spine (Figure 1). This is a result of the rich arterial blood supply of the thoracic spine and the blood reflux in the valveless Batson’s plexus in the lumbar spine. This pattern of distribution has been described by various other authors.3,6,12,13
Tuberculosis typically starts in the anterior and peri-discal areas,2 resulting in the typical radiographic image described earlier. Pande reported the incidence of atypical lesions to be 2.1%.14 He described these atypical lesions as a concentric collapse of a vertebra, an ivory vertebra, involvement of the neural arch, circumferential involvement of vertebrae and multiple vertebral diseases.
In this series, atypical lesions were found in 27% (12 of 43) of HIV-positive patients. There is no difference (p > 0.05) between HIV-negative and HIV-positive patients. In a smaller series of 30 patients, Sagane et al. also found no relation between the radiologic pattern and HIV status.5
Multiple non-contiguous spine lesions (skip lesions) are one of the atypical features of spinal TB.3,15 They were found in 10.53% of patients in this series: 6.5% were HIV-negative and 11.9% were HIV-positive. We found no statistically significant difference in the distribution of skip lesions between the two patient groups. Anley et al.4 and Sagane et al.5 also found no statistical difference in skip lesions between the two patient groups. Tuberculosis paralysis is, however, a devastating condition. This explains why several authors, even with fewer numbers of patients with skip lesions, recommend whole spine MRI when spinal TB is suspected.15,16,17 The patient in Figure 3 presenting with back pain and a thoracic gibbus illustrates this point. Plain thoracic spine radiographs showed a lesion consistent with TB. If whole spine MRI had not been performed on patients with suspected spinal TB, the skip lesion in the lumbar spine, as seen in the lumbar spine T1 Weighted Images (TWI) F/S post-contrast sequences, would have been missed. A biopsy of both thoracic and lumbar lesions confirmed TB.
There was no statistically significant difference between the two groups in the number of vertebrae involved (p = 0.07). The findings are similar to those of Sagane5 and Anley.4
There is no difference in kyphotic deformity between the two patient groups (p = 0.326). This is different to the findings of Sagane et al. and Anley et al., who found that there was a statistically greater degree of kyphosis in HIV-negative patients.4,5 This difference may be explained by the higher number of atypical vertebrae in the HIV-positive group in this study. As illustrated in Figure 2, atypical lesions do not show as much kyphosis as the typical spinal TB lesion. This may have a bearing on the extent of kyphosis in the HIV-positive patients in this series.
Anley et al. found no difference in the volume of pus between the two patient groups. Sagane et al. showed a significant statistical difference between the two (p < 0.0001), with HIV-positive patients depicting more pus formation. Both studies measured pus in the extradural space. This study measured the total volume of pus formed. Although not statistically significant, our results show a trend towards more pus formation in HIV-positive patients. This is likely related to the inflammatory response evoked by HIV co-infection.18
In the studies of Anley et al. and Sagane et al., vertebral body destruction was more significant in HIV-negative patients. In our study, vertebral body destruction, as measured by loss of vertebral body height, was greater in HIV-positive patients, and this is statistically significant (p = 0.01).
The increased inflammatory response seen in HIV-positive patients and resultant tissue necrosis is responsible for the tissue damage seen in spinal TB.19 Spinal TB is a paucibacillary disease. The extent of tissue destruction cannot be explained only by the effect of the few bacilli. The receptor activator of the nuclear factor-κ ligand (RANK-RANK-L) pathway is strongly activated in spinal TB.8,20 The activation of the RANK-RANK-L pathway is caused by cytokines, especially Tumour Necrosis Factor-alpha (TNF-α) and Interleukin (IL)-6.8 Histopathology studies report a higher expression of TNF-α in HIV-positive patients than in HIV-negative patients7; hence more bone destruction in HIV-positive patients. In clinical studies, the extent of destruction of pulmonary tissue is related to the expression of cytokines, particularly TNF-α.7 It is known that in severe pulmonary TB, there is an intense immune reaction leading to tissue destruction. The intense immune reaction in spinal TB, as found in the presence of HIV infection, also leads to bone destruction.8
These factors suggest that bone and tissue destruction is more pronounced in HIV-positive patients.8,20
Limitations
Only patients undergoing surgery were included in the study to ensure the availability of sufficient tissue for diagnostic confirmation. The results may, therefore, not be generally applicable to all patients with spinal TB.
Our HIV-positive patients have been adequately treated with antiretroviral drugs. This may affect the results.
Conclusion
Human immunodeficiency virus-positive patients exhibit more vertebral body destruction and a trend towards more pus formation in this study. This bone destruction is attributed to the activation of the RANK-L pathway by cytokines, a process more pronounced in HIV infection. Given the presence of skip lesions in spinal TB, we recommend that an MRI of the entire spine be performed when TB is suspected.
Acknowledgements
The authors acknowledge Sibongile Bopape for data capturing, Natasha Kleinhans for typesetting, and Paulinah Mhlanga for typing.
This article is based on research conducted as part of the PhD thesis of the corresponding author, Mthunzi Ngcelwane, entitled ‘Tuberculosis of the spine in HIV-negative and positive patients: Immune and pathogen-related factors in the local distribution of the disease’ submitted to the Department of Orthopaedics, Faculty of Health Sciences, University of Pretoria in 2022. The supervisors were Prof. Gregory Tintinger, Prof. Mishack Bida and Prof. Silas Motsitsi. The original thesis is available at: https://repository.up.ac.za/handle/2263/86036.
Part of this article is also based on an oral presentation to the 67th Congress of the South African Orthopaedic Association, held on 30 August 2021–02 September 2021. Refer: https://saoa.org.za/wp-content/uploads/2021/08/5.-SAOA-2021-Abstract-Book-26-Aug-2021.pdf.
Competing interests
The authors declare that they have no financial or personal relationships that may have inappropriately influenced them in writing this article.
Authors’ contributions
M.N. was responsible for conceptualisation, design, data acquisition, data analysis and administration. F.S. was responsible for radiographs and MRI analysis, and critical revision. N.M. carried out the design and critical revision. E.T. was responsible for radiograph analysis and data acquisition. M.M., L.B., A.Y., N.M. and J.S. carried out data acquisition. S.O. was responsible for statistical analysis. All authors read and approved the final manuscript.
Data availability
Data for this study are available from the corresponding author, N.M., and are kept at the Department of Orthopaedics, Steve Biko Academic Hospital, Pretoria.
Disclaimer
The views and opinions expressed in this article are those of the authors and are the product of professional research. They do not necessarily reflect the official policy or position of any affiliated institution, funder, agency or publisher. The authors are responsible for this article’s results, findings and content.
ReferencesWorld Health Organization. Global tuberculosis report 2020 [homepage on the Internet]. Geneva: World Health Organization; 2020[cited n.d.]. Available from: https://www.who.int/publications/i/item/9789240013131AnsariS, AmanullahF, AhmadK, RauniyarRJ. Pott’s spine: Diagnostic imaging modalities and technology advancements. N Am J Med Sci. 2013;5(7) 404–411. https://doi.org/10.4103/1947-2714.115775PandeKC, PandeSK, BabhulkarSS. An atypical presentation of tuberculosis of the spine. Spinal Cord. 1996;34(12):716–719. https://doi.org/10.1038/sc.1996.130AnleyCM, BrandtAD, DunnR. Magnetic resonance imaging findings in spinal tuberculosis: Comparison of HIV-positive and negative patients. Indian J Orthop. 2012;46(2):186–190. https://doi.org/10.4103/0019-5413.93688SaganeSS, PatilVS, BartakkeGD, KaleKY. Assessment of clinical and radiological parameters in spinal tuberculosis: Comparison between human immunodeficiency virus-positive and human immunodeficiency virus-negative patients. Asian Spine J. 2020;14(6):857–863. https://doi.org/10.31616/asj.2019.0251MaraisS, RoosI, MithaA, MabushaSJ, PatelV, BhigjeeAI. Spinal tuberculosis: Clinicoradiological findings in 274 patients. Clin Infect Dis. 2018;67(1):89–98. https://doi.org/10.1093/cid/ciy020IzawaK. Histological analysis of bone destruction in spinal tuberculosis. Kekkaku. 2015;90(3):415–420.BezuidenhoutJ, RobertsT, MullerL, Van HeldenP, WalzlG. Pleural tuberculosis in patients with early HIV infection is associated with increased TNF-alpha expression and necrosis in granulomas. PLoS One. 2009;4(1):e4228. https://doi.org/10.1371/journal.pone.0004228ShettyA, KannaRM, RajasekaranS. TB spine – Current aspects on clinical presentation, diagnosis, and management options. Semin Spine Surg. 2016;28(3):150–162. https://doi.org/10.1053/j.semss.2015.07.006RajasekaranS, SoundararajanDCR, ShettyAP, KannaRM. Spinal tuberculosis: Current concepts. Glob Spine J. 2018;8(4Suppl.):96S–108S. https://doi.org/10.1177/2192568218769053RajasekaranS, ShanmugasundaramTK. Prediction of angle of gibbus deformity in tuberculosis of the spine. J Bone Joint Surg Am. 1987;69-A(4):503–509. https://doi.org/10.2106/00004623-198769040-00005KandalaM, SugnaneswarP, VaishnaviV, et al. Clinico-radiological correlation with histopathology and molecular diagnosis in spinal tuberculosis. Int J Orthop. 2017;4(1):1–5.MohantySP, BhatS, NairSN. An analysis of clinicoradiological and histopathological correlation in tuberculosis of the spine. J Indian Med Assoc. 2011;109(3):161–165.PandeKC, BabhulkarSS. Atypical spinal tuberculosis. Clin Orth Relat Res. 2002;398:67–74. https://doi.org/10.1097/00003086-200205000-00010KannaRM, BabuN, KannanM, ShettyAP, RajasekaranS. Diagnostic accuracy of whole spine magnetic resonance imaging in spinal tuberculosis validated through tissue studies. Eur Spine J. 2019;28(12):3003–3010. https://doi.org/10.1007/s00586-019-06031-zWuM, SuJ, YanF, CaiL, DengZ. Skipped multifocal extensive spinal tuberculosis involving the whole spine: A case report and literature review. Medicine (Baltimore). 2018;97(3):e9692. https://doi.org/10.1097/MD.0000000000009692KailaR, MalhiAM, MahmoodB, SaifuddinA. The incidence of multiple level non-contiguous vertebral tuberculosis detected using whole spine MRI. J Spinal Disord Tech. 2007;20(1):78–82. https://doi.org/10.1097/01.bsd.0000211250.82823.0fChangCC, CraneM, ZhouJ, et al. HIV and co-infections. Immunol Rev. 2013;254(1):114–142. https://doi.org/10.1111/imr.12063MoonMS. Tuberculosis of spine – Contemporary thoughts on current issues and perspective views. Curr Orthop. 2007;21(5):364–379. https://doi.org/10.1016/j.cuor.2007.09.006NgcelwaneM. Tuberculosis of the spine in HIV-negative and -positive patients: Immune and pathogen related factors in the local distribution of the disease [homepage on the Internet] [doctoral thesis]. [cited n.d.]. Available from: https://repository.up.ac.za/handle/2263/86036
How to cite this article: Ngcelwane M, Maku M, Suleman F, et al. Bone destruction and radiologic features in surgically treated human immunodeficiency virus-negative and -positive patients with tuberculosis of the spine. J Coll Med S Afr. 2025;3(1), a151. https://doi.org/10.4102/jcmsa.v3i1.151