Vertebral osteomyelitis caused by nontuberculous mycobacteria: a case series of 23 patients and a review of 108 cases in the literature

Abstract

Background

Nontuberculous mycobacteria (NTM) are ubiquitous environmental opportunistic pathogens that rarely cause spinal infections. The clinical presentation, diagnostic process, and treatment strategies for NTM vertebral osteomyelitis often resemble those of spinal tuberculosis, posing significant challenges to clinical management.

Objective

This study aims to systematically analyze the clinical characteristics, imaging findings, microbial profiles, treatment approaches, and outcomes of NTM vertebral osteomyelitis, with a comparative review of relevant global literature to further elucidate their features.

Methods

We retrospectively analyzed 23 cases of NTM vertebral osteomyelitis diagnosed at our institution from 2021 to 2025. Inclusion criteria included spinal lesions with local and/or systemic symptoms, confirmed by positive NTM culture and/or molecular testing of biopsy or aspirate samples. Exclusion criteria were: (1) infections confirmed to be caused by Mycobacterium tuberculosis or other non-NTM pathogens, or cases with incomplete data; (2) patients with severe heart, liver, kidney dysfunction, or psychiatric disorders. Data collected included demographics, medical history, clinical presentation, imaging and laboratory findings (e.g., white blood cell count, C-reactive protein, erythrocyte sedimentation rate), microbiological profiles, and treatment outcomes. A literature review was conducted using PubMed, Google Scholar, and MEDLINE to retrieve studies on NTM spinal infections from 1970 to 2025 for comparative analysis.

Results

The study included 23 patients, with 43.5% (10/23) male and 56.5% (13/23) female, aged 38–80 years (mean: 57.35 ± 13.5 years). Predisposing factors included long-term glucocorticoid use (7/23), prior surgery (7/23), diabetes (7/23), and Human Immunodeficiency Virus (HIV) infection (2/23). All patients (100%, 23/23) presented with progressive spinal pain and restricted mobility, with fever in 5/23 (21.7%). Imaging revealed vertebral destruction in 87.0% (20/23) and intervertebral disc involvement in 52.2% (12/23). Microbiologically, Mycobacterium avium complex (MAC) was the most common pathogen (73.9%, 17/23), followed by Mycobacterium abscessus (M. abscessus) (13.0%, 3/23), with one case each of Mycobacterium iranicum (M.iranicum), Mycobacterium genavense (M. genavense), and Mycobacterium gordonae (M. gordonae). All patients received antibiotic therapy, with 78.3% (18/23) undergoing surgical intervention. After a mean follow-up of 18 ± 3.7 months (range, 12–24 months),13 patients showed complete or marked clinical improvement, Outcome data were not available for one patient who abandoned treatment. The literature review identified 92 studies reporting 108 NTM vertebral osteomyelitis cases.

Conclusion

NTM vertebral osteomyelitis is rare and diagnostically challenging due to their similarity to spinal tuberculosis. Early microbial identification is critical for individualized treatment. This case series of 23 patients significantly enriches the literature on NTM spinal infections, providing valuable clinical insights. Larger studies are needed to further characterize NTM infections.

Introduction

Nontuberculous mycobacteria (NTM), also known as atypical mycobacteria, encompass all Mycobacterium species excluding Mycobacterium tuberculosis and M. leprae. To date, over 190 NTM species have been identified [1] . These organisms are ubiquitous in the environment, found in water, soil, and aerosols, and have gained increasing clinical attention as opportunistic pathogens [2]. NTM infections predominantly affect immunocompromised individuals, such as those with prolonged glucocorticoid use or prior surgical history, while infections in immunocompetent individuals are rare. NTM primarily cause pulmonary infections, accounting for 80–90% of cases, followed by infections of lymph nodes, skin, and soft tissues. Extrapulmonary infections occur in approximately 0.6–9.0% of cases, with NTM-induced vertebral osteomyelitis being exceedingly rare [2–5]. The clinical, radiological, and pathological similarities between NTM vertebral osteomyelitis and spinal tuberculosis often lead to misdiagnosis or delayed treatment. Vertebral osteomyelitis, characterized by infection of the vertebrae, intervertebral discs, or paraspinal tissues, has seen a rising incidence in recent years, presenting significant diagnostic and therapeutic challenges [6, 7]. Existing literature on NTM vertebral osteomyelitis is limited, consisting mainly of case reports or small case series. This study analyzes 23 confirmed NTM vertebral osteomyelitis cases treated at our institution and reviews 108 well-documented cases from the global literature to explore the clinical characteristics, diagnostic approaches, treatment strategies, and outcomes of NTM spinal infections, aiming to provide evidence-based support for early diagnosis and management.

Materials and methods

Study design and patient selection

This retrospective single-center study included 23 patients diagnosed with nontuberculous mycobacterial (NTM) spinal infections at our institution, a specialized center for spinal infection management and the only dedicated spinal infection surgical department in China, between 2021 and 2025. Inclusion criteria were: (1) clinical symptoms of spinal pain and restricted mobility, with or without systemic symptoms (e.g., fever, weight loss); (2) imaging evidence of vertebral destruction, intervertebral disc infection, or involvement of spinal structures or adjacent soft tissues; and (3) microbiological confirmation of NTM infection via culture or gene sequencing with identified species. Exclusion criteria included: (1) spinal infections caused by Mycobacterium tuberculosis or other non-NTM pathogens; (2) incomplete clinical data or loss to follow-up. Surgical intervention was indicated for spinal instability, neural compression, epidural abscess formation, or progressive deterioration despite medical treatment, combined with anti-NTM therapy.

Data collection

Detailed data were collected on clinical symptoms (e.g., focal spinal pain, neurological deficits, fever), laboratory findings such as white blood cell(WBC)count, C-reactive protein (CRP), erythrocyte sedimentation rate (ESR),Mantoux test (MT), T-cell spot test (T-spot), imaging findings (e.g., T1-weighted imaging and T2-weighted imaging signals on MRI), acid-fast bacilli (AFB) staining, histopathological examinations, species identification and antimicrobial susceptibility.These data were analyzed to characterize the clinical features, diagnosis, treatment, and prognosis of NTM vertebral osteomyelitis.

For the included cases in our study, treatment outcomes were evaluated based on previous literature [8, 9] and clinical practice and were categorized into “recovery,” “improved,” and “treatment failure,” with predefined criteria for each category. Recovery was defined as complete or marked clinical improvement, characterized by substantial pain relief with marked improvement in VAS and ODI scores; normalization of inflammatory markers (CRP < 10 mg/L and ESR < 20 mm/h); and imaging stabilization or improvement, including resolution or reduction of abscesses, improvement of MRI signal abnormalities, and/or evidence of bone healing. Improved was defined as partial clinical improvement with pain relief and improvement in VAS and ODI scores; a decline in inflammatory markers; and imaging stabilization or improvement, manifested by partial resolution or reduction of abscesses and improvement of MRI signal abnormalities, with or without evidence of bone healing. Treatment failure was defined as no clinical improvement or worsening of symptoms, persistently unchanged or elevated inflammatory markers, and imaging progression, including enlargement or persistence of abscesses, worsening of MRI signal abnormalities, and absence of bone healing.

Lesion sampling, microbiological identification, and antimicrobial susceptibility testing

All patients with a confirmed diagnosis of non-tuberculous mycobacterial (NTM) infection underwent both lesion-based conventional culture and next-generation sequencing (NGS) on specimens obtained by CT-guided biopsy or lesion aspiration. All positive cultures were analyzed using matrix-assisted laser desorption/ionization time-of-flight mass spectrometry, which is the standard method in our laboratory for rapid and accurate mycobacterial speciation following growth on solid or liquid media. Next-generation sequencing (NGS) was performed directly on the original puncture specimens, without using cultured isolates. Nucleic acids were extracted from the specimens, followed by ultra-high-multiplex targeted amplification and sequencing on the Illumina platform. Sequencing data were then aligned against multi-level reference databases to achieve comprehensive pathogen identification. Performing NGS directly on clinical specimens allowed rapid detection, typically within 48 hours, including in culture-negative cases, and served as a complementary approach to conventional culture [3, 10, 11].

To minimize the risk of contamination, strict aseptic sampling procedures were followed, and negative controls were included in each NGS run. Isolates were classified as true pathogens based on multiple lines of evidence. Specifically, results from NGS and lesion cultures were interpreted in an integrated manner and correlated with the patients’ clinical presentation and radiologic features, and, when available, were further supported by histopathological findings.

Only pathogens with sufficient sequencing read depth and a clear correlation with the clinical and radiological findings were considered true positives, whereas potential background or contaminant organisms were excluded based on established laboratory thresholds and clinical judgment. In cases where microbiological results were inconclusive, or the likelihood of environmental contamination was considered high, repeat biopsy was performed and additional lesion specimens were submitted for NGS and/or culture to ensure diagnostic accuracy. For low-pathogenicity NTM species such as M. gordonae, classification as true pathogens rather than contaminants was based on a comprehensive assessment of clinical, radiological, and host factors, rather than on microbiological results alone. In all included cases, the diagnosis of spinal infection was strongly supported by clinical and imaging evidence, and NTM was the only pathogen identified, with no evidence of co‑infection with other bacteria.Moreover, patients with such low-virulence isolates demonstrated clinical and radiographic improvement in response to targeted anti-NTM therapy. In the single case of M. gordonae (Case 8), the patient had a history of prior lumbar spine surgery with retained internal fixation, which is a well-recognized risk factor for healthcare-associated NTM infection [12, 13]. Therefore,we consider the included isolates to most likely represent true infections rather than contaminants. Regarding diagnostic turnaround time, conventional culture results were typically available within approximately 30 days, whereas NGS results were obtained within 48–72 hours after specimen receipt.

Antimicrobial susceptibility testing was performed in accordance with the Clinical and Laboratory Standards Institute (CLSI) guidelines for non-tuberculous mycobacteria using the standardized broth microdilution method. Minimum inhibitory concentrations (MICs) were determined for relevant antimicrobial agents, and susceptibility categories (susceptible, intermediate, or resistant) were interpreted according to CLSI-recommended breakpoints. Susceptibility results were used to guide individualized antimicrobial regimens, with agents demonstrating in vitro susceptibility preferentially selected and those showing resistance avoided [14, 15].

Search strategy

A literature review was conducted using PubMed, Google Scholar, and MEDLINE databases. Search terms included “vertebral,” “spinal,” “infection,” “spondylodiscitis,” “discitis,” “osteomyelitis,” “atypical,” “nontuberculous,” and “mycobacterium,” combined with their MeSH terms and Boolean operators “AND” and “OR.” Case reports, case series, and review articles published between January 1970 and January 2025 were screened, and relevant data were extracted for comparative analysis. Following a PRISMA-inspired approach, the search yielded approximately 725 initial records. After removal of duplicates (n = 150) and screening of titles/abstracts (excluding n = 465 irrelevant records), 110 full-text articles were assessed for eligibility. Eighteen were excluded due to lack of confirmed NTM diagnosis or insufficient clinical data. Ultimately, 92 studies reporting 108 well-documented cases of NTM vertebral osteomyelitis were included for comparative analysis (Fig. 1).

Fig. 1.

PRISMA flow-chart

Statistical analysis

No statistical analysis was performed due to the descriptive nature of this case series.

Results

All patients underwent comprehensive diagnostic evaluations upon admission, including X-ray, CT, and full-spine MRI of the affected region, as well as laboratory tests (complete blood count, erythrocyte sedimentation rate）, CRP and blood cultures. Percutaneous biopsy or aspiration of the lesion site was performed, followed by routine anti-NTM treatment after confirmation of the causative pathogen via culture or next-generation sequencing. CRP and ESR were monitored regularly during treatment.

The study included 23 patients (10 males, 43.5%; 13 females, 56.5%), aged 38–80 years (mean: 57.35 ± 13.5 years). No cases were excluded. Lesion distribution included the lumbar spine (11/23), thoracic spine (8/23), and thoracolumbar junction (4/23). All patients (100%, 23/23) presented with focal spinal pain of varying severity and restricted mobility, with fever reported in 5/23 (21.7%). Among the 23 patients, 13 were lesion culture positive, 16 were NGS positive, with 6 patients being both positive. Among the 23 patients, those identified by NGS had no corresponding antimicrobial susceptibility results. We regret that, in some cases confirmed by lesion culture, further antimicrobial susceptibility testing was not performed, therefore, susceptibility results were available for only 9 patients. Of these, 3 patients exhibited macrolide resistance, and all were infected with MAC. In the patients with M. abscessus infection, susceptibility testing was available for only 1 case, which showed resistance to carbapenems and tetracyclines. Among the 9 patients with available antimicrobial susceptibility testing results, 8 were infected with MAC and 1 with M. abscessus. Testing was performed for multiple drug classes, including macrolides, carbapenems, tetracyclines, sulfonamides, aminoglycosides, oxazolidinones, quinolones, rifamycins, and first-line antituberculous agents, yielding susceptibility data for commonly used drugs such as clarithromycin, rifampicin, ethambutol, levofloxacin, and linezolid. The M. abscessus isolate (Case 1) was resistant to carbapenems and tetracyclines. Among the MAC isolates, resistance patterns were as follows: Case 3 was resistant to carbapenems, tetracyclines, and sulfonamides; Case 4 to carbapenems and tetracyclines; Case 6 to carbapenems, tetracyclines, and sulfonamides; Case 7 to macrolides, carbapenems, and tetracyclines; Case 10 to tetracyclines and sulfonamides; Case 15 to carbapenems, tetracyclines, and sulfonamides; Case 16 to tetracyclines and sulfonamides; and Case 19 to carbapenems, tetracyclines, and sulfonamides (Table 2).

Table 2.

Laboratory and microbiological characteristics of the 23 patients in this case series

Case	CRP(mg/L)	ESR(mm/h)	WBC(10 × 9/L)	drug sensitivity		MT	T-spot	Histopathologic	Acid fast staining
				Resistance Susceptible
1	24.24	56	3.81	CRO + Tet-R	Mac-S+ Sul-S+ Ozd-S+ Rif-S+ Fq-S+ Amg-S	-	-	without granuloma	/
2	26.33	80	6.18	/	/	-	-	Granuloma	/
3	27.77	20	4.09	CRO+ Tet-R+Sul-R	Mac-S + Ozd-S+ Rif-S+ Fq-S+ Amg-S	-	-	without granuloma	/
4	5.24	65	5.3	CRO + Tet-R	Mac-S + Sul-S+ Ozd-S+ Rif-S+ Fq-S+ Amg-S	-	-	Granuloma	-
5	13.94	45	5.43	/	/	+	-	/	/
6	43.98	88	5.45	CRO + Tet-R +Sul-R	Mac-S + Ozd-S+ Rif-S+ Fq-S+ Amg-S	/	-	Granuloma	+
7	3.21	14	4.33	Mac-R + CRO + Tet-R	Ozd-S+ Sul-S+ Rif-S+ Fq-S+ Amg-S	-	-	Granuloma	+
8	57.57	86	4.73	/	/	+	-	Granuloma	/
9	2.49	10	4.42	/	/	-	-	without granuloma	/
10	38.92	95	9.62	Tet-R + Sul-R	Mac-S+ Carb-S + Ozd-S+ Rif-S+ Fq-S+ Amg-S	/	-	Granuloma	-
11	88.23	72	7.94	/	/	/	-	without granuloma	/
12	91.98	40	7.34	/	/	-	-	/	/
13	53.63	30	4.12	/	/	-	-	without granuloma	/
14	68.06	45	4.8	/	/	/	-	without granuloma	+
15	65.53	38	4.88	CRO+ Tet-R+ Sul-R	Ozd-S+ Sul-S+ Rif-S+ Fq-S+ Amg-S	-	-	without granuloma	-
16	62.42	88	9.6	Tet-R + Sul-R	Mac-S+Carb-S+ Ozd-S+ Rif-S+ Fq-S+ Amg-S	-	-	without granuloma	-
17	33.04	46	4.56	/	/	/	+	without granuloma	/
18	57.18	28	18.2	/	/	+	+	Granuloma	/
19	61.2	95	11	CRO+ Tet-R +Sul-R	Mac-S+Ozd-S+ Rif-S+ Fq-S+ Amg-S	/	-	without granuloma	/
20	32.67	27	11.54	/	/	-	-	Granuloma	/
21	183.19	87	11.55	/	/	/	-	Granuloma	/
22	100.12	25	10.96	/	/	/	-	Granuloma	/
23	163.99	102	7.51	/	/	-	-	without granuloma	/

Note: Mac-R: Macrolide Resistance Mac-S: Macrolide-Susceptible CRO: Carbapenem Resistant Organisms Carb-S: Carbapenem-SusceptibleTet-R: Tetracycline Resistance Tet-S: Tetracycline Susceptible Sul-R: Sulfonamide Resistance Sul-S Sulfonamide Susceptible Amg-S: Aminoglycoside Susceptible Ozd-S: Oxazolidinone-Susceptible Rif-S: Rifamycin-Susceptible Fq-S: Fluoroquinolone-Susceptible S: Susceptible R: Resistant /: No relevant testing has been conducted

At admission, spinal nerve compression was observed in 5/23 patients, and spinal instability was noted in 18/23, with 18/23 undergoing surgical intervention. Five patients, lacking surgical indications, achieved resolution through conservative treatment. The time to diagnosis ranged from 3 to 40 days (Tables 1 and 2). Imaging revealed vertebral destruction in 87.0% (20/23) of cases, with some showing blurred vertebral margins and bone destruction. Intervertebral disc involvement was observed in 52.2% (12/23), with some cases exhibiting adjacent soft tissue involvement and spinal cord compression. We selected MRI and CT scans, pathological findings, and acid-fast staining results from a typical case, which revealed significant destruction of the L1 and L2 vertebral bodies, as well as the L1/L2 intervertebral disc, accompanied by large abscesses in both psoas muscles. Following anti-infective treatment, surgical intervention was carried out（Figs. 2, 3, and 4）.

Table 1.

Clinical characteristics of the 23 patients in this case series

Case	Age	Sex	Species	Spine involved	Site of infection	Presenting symptoms or signs
						Pain	Fever	Neurologic deficit
1	75	Female	Mycobacterium abscessus	thoracic	T8/9	yes	no	yes
2	75	Female	Mycobacterium avium complex	thoracolumbar	T12/L1	yes	no	no
3	58	Male	Mycobacterium avium complex	thoracic	T10/11、T6/7	yes	no	no
4	71	Female	Mycobacterium avium complex	lumbar	L3/4	yes	no	no
5	58	Male	Mycobacterium avium complex	thoracic	T5/6	yes	no	no
6	46	Female	Mycobacterium avium complex	thoracic	T10/11、T11/12	yes	no	no
7	59	Male	Mycobacterium avium complex	lumbar	L3/4	yes	no	no
8	51	Female	Mycobacterium gordonae	lumbar	L4/5、L5/S1	yes	no	yes
9	38	Female	Mycobacterium iranicum	lumbar	L4/5	yes	no	no
10	56	Female	Mycobacterium avium complex	thoracolumbar	T12/L1、L3/4	yes	yes	no
11	38	Male	Mycobacterium avium complex	lumbar	L2/3	yes	no	no
12	57	Male	Mycobacterium abscessus	thoracic	T1、T5、T6	yes	no	yes
13	42	Female	Mycobacterium avium complex	thoracic	T11/12	yes	yes	no
14	36	Female	Mycobacterium avium complex	lumbar	L1/2	yes	no	no
15	67	Male	Mycobacterium avium complex	thoracic	T10/11	yes	no	no
16	72	Male	Mycobacterium avium complex	thoracolumbar	T12/L1、L1/2	yes	yes	no
17	65	Male	Mycobacterium abscessus	lumbar	L4/5/S1	yes	no	no
18	61	Male	Mycobacterium avium complex	lumbar	L3/4/5	yes	no	yes
19	71	Female	Mycobacterium avium complex	thoracic	T11/12	yes	yes	yes
20	38	Female	Mycobacterium avium complex	thoracolumbar	T12/L1	yes	no	no
21	41	Female	Mycobacterium genavense	lumbar	L1、L2、L3、L4、L5	yes	no	no
22	64	Female	Mycobacterium avium complex	lumbar	L1/2	yes	yes	no
23	80	Male	Mycobacterium avium complex	lumbar	L2/3	yes	no	no

Note: MT: Mantoux test; T-spot: T-cell Spot Test;-: Negative; +: positive;/:no record

Fig. 2.

Case 14 at admission: lumbar T2-weighted fat-suppressed and T1-weighted contrast-enhanced MRI images demonstrating L1 and L2 vertebral destruction, accompanied by an intraspinal abscess and bilateral psoas abscesses

Fig. 3.

Case 14 after anti-NTM treatment: (A, B) preoperative and postoperative CT images illustrating changes in bilateral psoas abscesses. (C, D) postoperative X-ray images. There is a noticeable decrease in the abscess

Fig. 4.

A shows granulomatous inflammation, and B shows positive acid-fast bacilli staining (arrows indicate the organisms), which is histologically indistinguishable from tuberculous infection

Laboratory findings included elevated WBC counts in 4/23 patients (17.4%), elevated ESR in 21/23 (91.3%), and elevated CRP in 20/23 (87.0%). T-spot testing was positive in 2/23 (8.7%), and the Mantoux test was positive in 3/23 (13.0%). Histopathology revealed granulomatous inflammation in 6/23 (26.1%) cases, with acid-fast bacilli staining positive in 3/23 (13.0%) (Table 1). All cases were confirmed as NTM infections via tissue biopsy culture or NGS. Microbiological profiling identified MAC as the predominant pathogen (73.9%, 17/23), followed by M. abscessus (13.0%, 3/23), with one case each of M. iranicum, M. genavense, and M. gordonae. Notably, M. iranicum, a rare species, was detected in one patient with pronounced localized vertebral destruction but mild clinical symptoms, suggesting potential differences in pathogenicity and clinical presentation among NTM species. In one case, a 36-year-old female presented with L1 and L2 vertebral and disc destruction and bilateral psoas abscesses. CT-guided biopsy of the L1/2 disc confirmed MAC infection, with histopathology showing granulomatous inflammation, inflammatory necrosis, and minimal dead bone; acid-fast bacilli staining was positive. Among the 23 cases, 6/23 (26.1%) showed granulomatous inflammation on histopathology, with only 3/23 (13.0%) positive for acid-fast bacilli staining.

Discussion

NTM are ubiquitous in environmental sources such as water, dust, and soil, with transmission typically linked to contaminated water, soil, or aerosols. While pulmonary infections are the most common manifestation of NTM, spinal infections are exceedingly rare. The pathogenesis of NTM infections remains incompletely understood, but current evidence suggests an interplay of host, environmental, and bacterial factors, creating a “susceptible host” profile. Immunodeficiency, prolonged use of corticosteroids, or immunosuppressive therapies significantly increases the risk of NTM infection, particularly in patients with systemic lupus erythematosus or HIV [1, 16, 17]. In immunocompetent individuals, NTM infections involving the skin, soft tissues, joints, or bones are typically attributed to direct inoculation following penetrating trauma or contamination during surgical or invasive procedures, such as animal bites, injections, cardiac surgery, or fracture repair [18–20]. The first documented case of NTM osteomyelitis was reported by Weed et al. in 1956 [21], highlighting the long-recognized but uncommon nature of this entity.In the present cohort, 39.1% (9/23) of patients had underlying immunodeficiency or were receiving immunosuppressive therapy. Similarly, among the 108 reported cases of NTM vertebral osteomyelitis identified in the literature, 40.7% (44/108) occurred in immunocompromised individuals, including 22.2% (24/108) with autoimmune diseases receiving long-term corticosteroid therapy, 13.0% (14/108) with HIV infection, and 5.6% (6/108) with congenital immunodeficiencies. Collectively, these findings underscore immunodeficiency as a major predisposing factor for NTM vertebral osteomyelitis. In our cohort, seven patients had a history of prior spinal surgery. Among these cases, MAC was identified in four patients, M. abscessus in two, and M. gordonae in one. Overall, immunosuppression defined as long-term glucocorticoid use (n = 7) or HIV infection (n = 2) was present in nine patients, among whom MAC predominated (8/9 cases), with M. genavense identified in the remaining case. These observations are partially consistent with previous reports [20, 22–24], which suggest that MAC is more frequently encountered in immunocompromised hosts, whereas rapidly growing mycobacteria, such as M. abscessus, are often associated with healthcare-related or post-procedural infections. Nevertheless, given the relatively small sample size and the limited number of cases within each subgroup, definitive conclusions regarding species-specific risk factor predilections cannot be drawn. Larger, multicenter studies are therefore warranted to further clarify these associations.

Similar to soft tissue or non-spinal bone infections, NTM vertebral osteomyelitis is often associated with trauma, penetrating injuries, injections, or prior surgery. In this study, 4/23 patients had infections at sites of prior surgical intervention, while cases without evident trauma or surgical history likely resulted from hematogenous spread, particularly in patients with underlying immunosuppression. Spinal tuberculosis is thought to arise from hematogenous dissemination following pulmonary or extrapulmonary infection, starting in subchondral bone and spreading to the intervertebral disc and adjacent vertebrae. NTM vertebral osteomyelitis may follow a similar pathway, with pathogens transported via macrophages to Batson’s venous plexus, leading to anterior spinal involvement. The limited vascular supply to vertebral laminae and discs may delay early destruction [25, 26].

All 23 patients in this study presented with varying degrees of low back or thoracic pain, predominantly at night, with 11/23 involving the lumbar spine, 8/23 the thoracic spine, and 4/23 the thoracolumbar junction. Fever was noted in 5/23 patients, and 2/23 developed lower limb weakness due to abscess-related compression. Single-segment involvement occurred in 60.9% (14/23) of cases. In the literature, 90.7% (98/108) of patients reported back pain, 36.1% (39/108) had lower limb weakness, with thoracic involvement in 56.5% (61/108) and lumbar in 41.7% (45/108) (Table 3). Both NTM vertebral osteomyelitis and tuberculous spondylitis present with granulomatous inflammation, often leading to misdiagnosis as tuberculosis. NTM vertebral osteomyelitis typically shows mild paraspinal soft tissue swelling, with rare, thin paravertebral abscesses that may contain calcifications and are confined to adjacent vertebrae. In contrast, tuberculous spondylitis is characterized by large, calcified paravertebral abscesses extending beyond the affected vertebrae, with less frequent extraspinal skeletal or systemic soft tissue involvement.An extensive paraspinal abscess, which exceeds the length of the affected vertebral body and is frequently accompanied by calcification, is a typical manifestation of spinal tuberculosis. Additionally, it rarely involves bone involvement outside the spine or systemic soft tissue abscesses [117].

Table 3.

Clinical characteristics, management, and outcomes of patients with NTM-induced vertebral osteomyelitis （not including current series）108 [23, 24, 27–116]

Category	Feature	Data
Demographic characteristics	Mean age（years）	51.5 ± 17.3
	<18 years	4 (3.7%)
	>65 years	39(36.1%)
	Male	61 (56.5%)
Infection characteristics	Spinal involvement site
	Thoracic spine	61 (56.5%)
	Lumbar spine	45 (41.7%)
	Cervical spine	1 (0.9%)
	Entire spine	4 (3.7%)
	Involvement pattern
	Single segment	74 (68.5%)
	Multiple segments (including entire spine)	34 (31.5%)
	Abscess complications	81(75.0%)
	Paraspinal abscess	45 (41.7%)
	Epidural abscess	14(12.9%)
	Psoas abscess	4 (3.7%)
	Multiple abscesses	17 (15.7%)
	Extraspinal dissemination	30 (27.8%)
	Pulmonary	9 (8.3%)
	Bone infection at other sites	14 (13.0%)
	Multiple dissemination	9 (8.3%)
Clinical symptoms	Back pain	98 (90.7%)
	Lower limb weakness	39 (36.1%)
	Fever	31 (28.7%)
Immune status	Intact immune function	64 (59.3%)
	Impaired immune function	44 (40.7%)
	Autoimmune disease + immunosuppression	24（22.2%）
	HIV infection	14 (13.0%)
	Congenital immunodeficiency	6 (5.6%)
Predisposing factors (immunocompetent)	Spinal surgery/epidural injection	10 (9.3%)
	Previous tuberculosis infection	6 (5.5%)
	Diabetes	6 (5.5%)
Diagnostic methods	Percutaneous needle biopsy	51 (47.2%)
	Intraoperative lesion sampling	38 (35.2%)
	Sputum/abscess drainage/blood culture	19 (17.6%)
Microbiology	Species distribution
	Mycobacterium avium complex	46 (42.6%)
	Mycobacterium xenopi	14 (13.0%)
	Mycobacterium abscessus	9 (8.3%)
	Mycobacterium chelonae	8 (7.4%)
	Mycobacterium kansasii	8 (7.4%)
	Mycobacterium fortuitum	7 (6.5%)
	Mycobacterium chimaera	3 (2.7%)
	Mycobacterium simiae	2 (1.9%)
	Other 11 rare species	11 (10.2%)
	Confirmation method
	Culture	76 (70.4%)
	PCR/NGS	32 (29.6%)
Treatment	Treatment strategy
	Antibiotics alone	45 (41.7%)
	Surgery + antibiotics	63 (58.3%)
	Surgical approach (n = 58)
	Posterior decompression	14 (24.1%)
	Abdominal/psoas drainage	9 (15.5%)
	Anterior decompression and fusion	8 (13.9%)
	Combined anterior-posterior fusion	12 (20.7%)
	Posterior decompression and fusion	15 (25.9%)
	Commonly used drugs
	Ethambutol	67 (62.0%)
	Rifampicin	60 (55.6%)
	Clarithromycin	52 (48.1%)
Prognosis	Clinical outcome (n = 93)
	Recovery	31 (33.3%)
	improved	45 (48.4%)
	Treatment failure	8 (8.6%)
	Death	9 (9.7%)

Notes: Other 11 rare species (each 1 case): Mycobacterium scrofulaceum, Mycobacterium smegmatis, Mycobacterium malmoense, Mycobacterium phlei, Mycobacterium heckeshornense, Mycobacterium flavescens, Mycobacterium arosiense，Mycobacterium gordonae，Mycobacterium paragordonae，Mycobacterium monacense, Mycobacterium chelonae

Diagnosing NTM vertebral osteomyelitis is challenging due to its insidious onset, slow progression, and non-specific symptoms, which resemble those of tuberculosis or low-virulence bacterial infections, often delaying diagnosis. Common symptoms include chronic back pain, localized tenderness, restricted mobility, low-grade fever, night sweats, fatigue, and weight loss, which are indistinguishable from tuberculosis. Imaging findings lack specificity, showing vertebral destruction, disc space narrowing, or paravertebral abscesses similar to tuberculosis, or rapid bone destruction resembling pyogenic infections [118]. NTM cultures are difficult and time-consuming, often requiring up to a month for results, with low positivity rates, especially in blood cultures, due to contamination risks or misidentification as commensal bacteria. Slow-growing NTM require specialized media, optimal temperature, and CO₂ concentrations for culture [3,110]. In this study, one patient was misdiagnosed for 40 days and treated for tuberculosis before NTM confirmation. Among the 23 patients, CRP and ESR were elevated to varying degrees, but WBC counts were normal in 16/23. T-spot and Mantoux tests were positive in only 2/23 and 3/23 patients, respectively, indicating limited diagnostic utility for NTM vertebral osteomyelitis.

Imaging for NTM vertebral osteomyelitis shows characteristic changes on plain radiographs 3–4 weeks after disease onset, with more pronounced changes after approximately 2 months, though spinal deformity is rare and non-specific. CT and MRI detect changes earlier and with greater specificity, revealing paraspinal soft tissue swelling and vertebral destruction with mixed sclerotic rims, distinguishing NTM vertebral osteomyelitis from tuberculous or pyogenic spondylitis, where tuberculosis shows prominent sequestra and abscesses, and pyogenic infections lack significant osteogenesis [119, 120]. Smimmo et al. [106] noted that NTM spondylitis may involve multiple vertebrae or the entire spine, with multiple central or marginal lytic lesions within a single vertebra and frequent appendage involvement, reflecting its propensity for dissemination. Tuberculous spondylitis typically affects adjacent vertebrae with severe bone destruction and sequestra, predominantly marginal lesions, and rarely multiple lesions within a single vertebra. NTM vertebral osteomyelitis exhibits both osteoblastic and patchy osteolytic destruction, with minimal psoas abscesses and possible multi-segmental or skip lesions, likely due to NTM’s lower pathogenicity and prolonged disease course [25, 102]. In this study, 87.0% (20/23) of cases showed vertebral destruction, some with blurred margins, and 52.2% (12/23) had disc involvement, with some exhibiting adjacent soft tissue involvement and spinal cord compression.

Microbiological diagnosis of NTM vertebral osteomyelitis requires NTM isolation from spinal lesions (vertebrae, discs, or abscess fluid) via culture and/or histopathological evidence of granulomatous inflammation with positive acid-fast bacilli staining and molecular confirmation identifying NTM. Typical histopathology shows granulomatous inflammation (epithelioid cells, lymphocytic infiltration) with possible caseating or non-caseating necrosis. Acid-fast bacilli staining may be positive but is less sensitive than culture, and histopathological findings must be corroborated by microbiological results to differentiate NTM from tuberculosis [4]. In this study, all 23 cases underwent both conventional lesion culture and next-generation sequencing (NGS) on specimens obtained by CT-guided biopsy or aspiration. Culture was positive for NTM in 13 of 23 cases, NGS yielded a definitive NTM identification in 16 of 23 cases, and 6 cases were positive by both methods. NGS provided results within 48 hours, compared to nearly a month for culture, highlighting its critical role in rapid diagnosis and timely treatment.

No standardized treatment guidelines exist for NTM vertebral osteomyelitis, with protocols often adapted from pulmonary NTM treatment. MAC infections, the most common in this study, are typically treated with clarithromycin or azithromycin, ethambutol, and rifampicin, with amikacin or streptomycin added for severe cases. In contrast, M. abscessus is resistant to first-line antituberculous drugs, requiring oral macrolides combined with intravenous amikacin, tigecycline, or imipenem [3, 121]. Treatment regimens vary by species. In this study, 7/23 patients received a quadruple regimen of clarithromycin, rifampicin, ethambutol, and moxifloxacin; 21/23 used clarithromycin, 19/23 rifampicin, and 16/23 ethambutol (Table 4), consistent with literature trends where ethambutol, rifampicin, and clarithromycin are the most frequently used drugs. One patient discontinued treatment due to heart and liver failure, while the remaining 22/23 showed varying degrees of improvement. Treatment regimens were individualized based on available susceptibility results and species-specific recommendations. Macrolides (clarithromycin or azithromycin) were included as cornerstone agents for macrolide-susceptible isolates (primarily MAC). For the M. abscessus case with documented inducible macrolide resistance and carbapenem/tetracycline resistance, the regimen avoided reliance on these classes and incorporated intravenous amikacin and cefoxitin, combined with oral linezolid, in line with CLSI-guided alternatives for resistant rapid growers.

Table 4.

Clinical characteristics, diagnosis, treatment, and outcome of 23 patients with vertebral osteomyelitis caused by Nontuberculous Mycobacteria

Case	Predisposing factors	Underlying diseases	Antibiotic regimen	Confirmation method	Time to species identification (days)	surgery	Outcome	Follow-up (months)
1	After percutaneous vertebro plasty surgery	None	CCR, A, CFX, MXF	Culture	6	yes	Recovery	15
2	Following bronchial artery embolization	Bronchiectasis complicated by hemoptysis	CCR, LZD, E, Le	NGS	2	yes	Improved	18
3	Long-term glucocorticoid use	systemic lupus erythematosus	CCR, R, MXF	Culture	24	yes	Improved	21
4	None	None	CCR, R, E, I	Culture	24	yes	Improved	12
5	None	None	CCR, R, E, Le	Culture	12	no	Recovery	24
6	None	None	CCR, R, E, MXF	Culture/NGS	40/2	yes	Recovery	15
7	None	Nontuberculous Mycobacterial Pulmonary Disease	R, LZD, E, MXF	Culture	7	yes	Recovery	21
8	A history of lumbar spine surgery	Diabetes	CCR, R, E, MXF	NGS	2	yes	Recovery	15
9	None	Diabetes	CCR, A, MC	NGS	2	no	Recovery	21
10	A history of lumbar spine surgery	Diabetes、Leukemia	CCR, R, LZD, A	Culture	11	yes	Improved	18
11	None	HIV	CCR, R, E, MXF	Culture/NGS	32/2	yes	Improved	15
12	Postoperative femoral fracture	None	CCR, R, E, MXF	NGS	2	no	Recovery	21
13	Long-term surgery use	Dermatomyositis、Diabetes	CCR, Rt, E, Le	Culture/NGS	Confirmed outside the hospital	yes	Improved	12
14	Long-term glucocorticoid use	systemic lupus erythematosus	CCR, R, E, Le	Culture/NGS	12/2	yes	Recovery	15
15	None	Diabetes、Hypertension、COPD	CCR, R, LZD, MXF	Culture/NGS	29/2	yes	Improved	21
16	None	None	CCR, R, E, MXF	Culture	33	no	Improved	24
17	Long term radiotherapy	lymphoma、Hypertension、Cerebral infarction	IPM, CFX, R, I, Le	NGS	2	yes	Improved	15
18	A history of lumbar spine surgery	None	CCR, R, E, MXF	NGS	2	yes	Recovery	21
19	Long-term glucocorticoid use	Scleroderma、Nontuberculous Mycobacterial Pulmonary Disease、Interstitial pneumonia、Hypertension	CCR, Rt, E, A	Culture+/NGS	Confirmed outside the hospital	yes	Recovery	15
20	Long-term glucocorticoid use	systemic lupus erythematosus	CCR, Rfb, E, A	NGS	Confirmed outside the hospital	yes	Recovery	21
21	None	HIV	R, E, MXF, A	NGS	2	yes	Recovery	15
22	Long-term glucocorticoid use、A history of lumbar spine surgery	Hypersensitivity Syndrome、Hypertension	CCR, R, E	NGS	2	yes	Recovery	21
23	Long-term glucocorticoid use、A history of lumbar spine surgery	Rheumatoid arthritis	CCR, R, E, A	NGS	2	no	Not available	Not available

Note: CCR: clarithromycin Rfb: rifabutin R: rifampicin I: isoniazid, E: ethambutol, MXF: moxifloxacin Le: levofloxacin A: amikacin, Rt: rifapentine  IPM: imipenem CFX: cefoxitin LZD: linezolid， MC: minocycline AHA: autoimmune hemolytic anemia; COPD: chronic obstructive pulmonary disease; DM: Diabetes mellitus; ILD: interstitial lung disease; SU: steroid use. Time to species identification: the duration from sample collection to definitive species identification

Surgical indications for NTM spinal infections lack consensus, but extensive lesions, abscess formation, progressive vertebral destruction, neurological deficits, or poor response to medical therapy warrant consideration [3,86,107]. Of the 23 patients in this study, 18 underwent surgery, including anterior or posterior lesion debridement, segmental fixation, and abscess drainage, and most of these patients achieved satisfactory clinical outcomes. Nevertheless, antimicrobial susceptibility testing was not available for all patients, with results obtained in only 9 of the 23 cases. Moreover, as a retrospective case series with a relatively small sample size, this study is subject to inherent reporting bias and limited generalizability. Accordingly, our therapeutic observations should be interpreted as preliminary rather than definitive, and large-scale, prospective studies are needed to establish and validate standardized treatment strategies.

Conclusion

Non-tuberculous mycobacterial spinal infections are rare and diagnostically challenging. Diagnosis generally relies on a combination of imaging findings, microbiological testing, and histopathological evaluation. Early pathogen identification may facilitate more tailored antimicrobial management. However, further larger studies are needed to better characterize this entity.

Author contributions

WR: collected the patient data and drafted the manuscript. RD and YY:helped collect the clinical and radiographic data of the patient. JL, LX, WZ,HA,JS: helped revise the draft. HT: performed the surgery and revised the manuscript critically. All authors read and approved the final manuscript.

Funding

None.

Data availability

The data are available from the corresponding authors upon reasonable request.

Declarations

Ethics approval and consent to participate

This study adhered to the principles outlined in the Declaration of Helsinki and was approved by the Ethics Committee of Shandong Public Health Clinical Center. Written informed consent was obtained from all the participants.

Consent for publication

Written informed consent was obtained from all the patients for publication of their information and accompanying images.

Competing interests

The authors declare no competing interests.