Differential analysis of the lower respiratory tract microbiota between patients with non-tuberculous mycobacterial pulmonary disease and pulmonary tuberculosis

Abstract

Background

Non-tuberculous mycobacterial pulmonary disease (NTM-PD) poses significant diagnostic challenges due to its clinical similarity to pulmonary tuberculosis (PTB). This study aimed to compare the clinical characteristics, immune status and lower respiratory tract microbiome profiles of NTM-PD and PTB patients.

Methods

A total of 113 NTM-PD patients and 105 PTB patients were enrolled. The clinical features, laboratory parameters, and comorbidity profiles were analyzed. Bronchoalveolar lavage fluid (BALF) samples were subjected to bacterial culture and targeted next-generation sequencing (tNGS) for microbiome characterization. Microbiome distributions were then compared across disease groups, host characteristics, immune cell subsets, and NTM species.

Results

Compared with PTB patients, NTM-PD patients were older and exhibited lower rates of smoking, alcohol use, interferon-gamma release assay (IGRA) positivity, red blood cell (RBC) count, hemoglobin (Hb) and albumin levels, but a higher erythrocyte sedimentation rate (ESR). Hemoptysis, bronchiectasis, and chronic obstructive pulmonary disease (COPD) were more common among NTM-PD patients, while lymphadenopathy and diabetes were more frequent among PTB patients. BALF microbiome analysis revealed distinct profiles: Pseudomonas aeruginosa and Aspergillus spp. were more frequently detected in NTM-PD patients, while Neisseria spp. and Streptococcus viridans predominated in PTB patients. Reduced CD4⁺ T cell counts were associated with a higher detection rate of Pseudomonas aeruginosa and Candida spp. in NTM-PD patients. The most prevalent NTM species were the M. avium complex (MAC) (62.37%) and the M. abscessus complex (MABC) (27.96%).

Conclusions

The lower respiratory tract microbiome in NTM-PD differs from that in PTB and is characterized by a high prevalence of Aspergillus spp. and Pseudomonas aeruginosa. These differences are likely influenced by underlying structural lung disease, host immune status, and the NTM species themselves. Our findings provide new insights for the precise diagnosis and treatment of NTM-PD.

Supplementary Information

The online version contains supplementary material available at https://doi.org/10.1186/s12866-026-05300-4.

Introduction

Non-tuberculous mycobacteria (NTM) are ubiquitous environmental opportunistic pathogens responsible for non-tuberculous mycobacterial pulmonary disease (NTM-PD). The global incidence and prevalence of NTM-PD have risen markedly in recent decades, representing an increasing public health challenge [1]. In China, epidemiological surveys over the past forty years have documented a consistent increase in NTM infection rates, a trend linked to rapid population aging, a rising burden of structural lung diseases, growing immunosuppressed populations, and complex pulmonary microbial colonization patterns [2, 3]. Clinically, NTM-PD shares substantial overlap with pulmonary tuberculosis (PTB) in terms of symptoms, radiographic features and laboratory findings, which often leads to delayed diagnosis and inappropriate therapeutic decisions [4]. However, these two diseases differ fundamentally in microbial biology, host susceptibility, management strategies, and clinical outcomes. Despite these critical distinctions, systematic comparative studies focusing on host immune status, lower respiratory tract microbiome composition, and host-microbe interactions in NTM-PD versus PTB remain scarce. A deeper understanding of these micro-level differences is essential for improving diagnostic precision and guiding therapeutic strategies.

Current differentiation between NTM-PD and PTB relies primarily on mycobacterial culture, traditional phenotypic identification and molecular species identification techniques [5]. However, these methods provide limited insight into the immunological and microbial ecological factors underlying susceptibility and disease progression. Although the interferon-gamma release assay (IGRA) is theoretically useful for distinguishing Mycobacterium tuberculosis (Mtb) infection from NTM infection, it is prone to false-positive or false-negative results, and its diagnostic value is still being debated [6].

Accumulating evidence indicates that the lower respiratory tract microbiome plays a pivotal role in the development, progression and immune regulation of chronic respiratory diseases [7–10]. Changes in airway microbial communities are closely linked to inflammation, disease chronicity, and treatment responsiveness. For example, specific microbial signatures are associated with exacerbation frequency and severity in bronchiectasis (BE) and chronic obstructive pulmonary disease (COPD) [11, 12]. Notably, BE and COPD are also common underlying conditions in NTM-PD patients, suggesting that airway dysbiosis may facilitate NTM colonization and infection [13, 14]. While prior studies have characterized respiratory microbiome alterations of PTB patients relative to healthy individuals, data on the microbiome characteristics of NTM-PD patients remain scarce [15]. Even fewer studies have systematically compared NTM-PD and PTB or explored associations between microbiome patterns, clinical features, and host immune function. In China, most existing research has focused on clinical features and pathogen detection, leaving significant gaps in our understanding of immune dysregulation, microecological disturbances and species-specific microbial patterns associated with different NTM pathogens [4, 16–18]. However, with the increasing application of metagenomic next-generation sequencing (mNGS) and targeted NGS (tNGS) in respiratory infection diagnostics, it has become feasible to profile airway microbial communities in high resolution [19, 20]. This provides a new opportunity to elucidate the complex interplay between the host, pathogen and microbiome.

Against this background, the present study compared the clinical characteristics, immune profiles and bronchoalveolar lavage fluid (BALF) microbiome compositions of patients with NTM-PD and PTB. The aim was to delineate differences in host status and microecological features between the two diseases, explore potential mechanisms linking microbial dysbiosis with disease phenotype and co-infection patterns, and provide evidence to support more accurate differential diagnosis and optimized treatment strategies.

Materials and methods

This study was designed to compare the differences in the lower respiratory tract microbiota between NTM-PD and PTB patients. The study comprised three main components: (1) Clinical laboratory tests: blood routine and biochemical examination; (2) Etiological examination: smear microscopy, culture, PCR, tNGS. (3) Immunological tests: lymphocyte subset analysis, IGRA, TB-antibody test. A detailed flowchart of the study design is shown in Fig. 1.

Fig. 1.

The detailed design diagram of this study. Venous blood samples from all the subjects were collected and analyzed separately for (1) laboratory tests and (3) immunological tests. Bronchoscopy was performed to instill and aspirate sterile saline from the bronchial alveoli to collect BALF for subsequent (2) etiological examination, including smear microscopy, culture, and PCR. Additionally, the BALF samples were sent to a third-party testing company for tNGS. Abbreviations: NTM-PD, non-tuberculous mycobacterial pulmonary disease; PTB, pulmonary tuberculosis; BALF, bronchoalveolar lavage fluid; PNB/TCH, paranitrobenzoic acid/2-thiophenecarboxylic acid hydrazide; tNGS, targeted next-generation sequencing; IGRA, interferon-gamma release assay; WBC, white blood cell count; RBC, red blood cell count; Hb, hemoglobin; PC, platelet count; ANC, absolute neutrophil count; ALC, absolute lymphocyte count; AMC, absolute monocyte count; ESR, erythrocyte sedimentation rate; hs-CRP, high-sensitivity C-reactive protein; PT, prothrombin time; APTT, activated partial thromboplastin time; Alb, albumin; PCT, procalcitonin

Subjects and ethics statement

This retrospective study was conducted among hospitalized patients and was approved by the Ethics Committee of the Eighth Medical Center of the PLA General Hospital (approval number: 309202109061503).

A total of 113 hospitalized patients diagnosed with NTM-PD between January 2020 and June 2025 were included. Diagnosis of NTM-PD was carried out according to the “Guidelines for the Diagnosis and Management of NTM Diseases” issued by the Chinese Society of Tuberculosis, Chinese Medical Association [21]. The diagnostic criteria included respiratory or systemic symptoms alongside chest imaging findings (e.g., cavitary shadows, multifocal BE, or small nodular lesions) after excluding other pulmonary diseases. A definitive diagnosis required meeting at least one of the following criteria: (1) Two separate sputum samples testing positive for the same NTM pathogen by culture and/or molecular testing; (2) One positive NTM culture and/or molecular test from bronchial washings or BALF; or (3) Lung biopsy demonstrating mycobacterial histopathological features accompanied by a positive NTM culture or molecular test from sputum or BALF.

For comparison, 105 PTB patients admitted during the same period were included. PTB was diagnosed in accordance with the “Diagnosis for Pulmonary Tuberculosis (WS288-2017)” issued by the National Health and Family Planning Commission of the People′s Republic of China [22].

Patients were excluded if they had any of the following: (1) concurrent NTM-PD and PTB; (2) autoimmune diseases; (3) severe hepatic or renal dysfunction; (4) HIV infection; or (5) pregnancy or lactation. These exclusion criteria were applied to minimize potential confounding factors that could affect host immune responses and respiratory microbiome composition independently. It is known that autoimmune diseases, HIV infection, and severe hepatic or renal dysfunction alter immune status, systemic inflammation, and microbial community structure, which could potentially bias the analysis of immune–microbiome interactions.

Clinical laboratory tests

Following an overnight fast, venous blood samples were collected from each subject for routine testing, including: white blood cell count (WBC), red blood cell count (RBC), hemoglobin (Hb), platelet count (PC), absolute neutrophil count (ANC), absolute lymphocyte count (ALC), absolute monocyte count (AMC), erythrocyte sedimentation rate (ESR), high-sensitivity C-reactive protein (hs-CRP), prothrombin time (PT), activated partial thromboplastin time (APTT), albumin (Alb), and procalcitonin (PCT). All analyses were performed by the hospital’s clinical laboratory.

Etiological examination

BALF and sputum samples were collected from all patients for etiological testing.

Smear microscopy was performed using Ziehl-Neelsen (Z-N) acid-fast staining. Mycobacterium culture was conducted using the BACTEC MGIT 960 system (BD Biosciences, Franklin Lakes, USA), with preliminary NTM species identification using PNB/TCH medium. For routine bacterial culture, samples were inoculated onto Columbia blood agar and Sabouraud agar (Autobio-diagnostics, Zhengzhou, China) and incubated at 37 ℃ for 18–24 h. Suspected colonies underwent Gram staining and microscopic examination; pure colonies were identified using the VITEK 2 Compact System (bioMérieux, France). Molecular testing utilized the Xpert MTB/RIF assay (Cepheid, Sunnyvale, CA, USA) and CapitalBio Mycobacterium nucleic acid detection (CapitalBio, China).

With informed consent, 3 ml of BALF from each patient was sent to a third-party laboratory (Hangzhou ShengTing Medical Technology Co., China) for pathogen detection via tNGS. In brief, total nucleic acids were extracted following enzymatic lysis and subjected to multiplex PCR amplification for target enrichment. Sequencing libraries were prepared through fragmentation, end repair, A-tailing, adapter ligation, and magnetic bead purification. These libraries were then sequenced on the GridION platform (Oxford Nanopore Technologies). The raw sequencing data were quality-filtered and analyzed using customized bioinformatics pipelines for pathogen identification. Only samples that met the predefined quality control criteria for nucleic acid purity (A260/A230 > 2.0 and 1.7 ≤ A260/A280 ≤ 1.9) were included in the analysis.

Immunological tests

Flow cytometry

A total of 4 ml of peripheral venous blood was collected from fasting participant into heparinized anticoagulant tubes, and stored at room temperature. Lymphocyte subsets were analyzed by flow cytometry as previously described [23]. Reference ranges were based on established normal values for the Chinese Han population [24]. CD3⁺: 834-2,217 cells/µL, CD4⁺: 395-1,264 cells/µL, CD8⁺: 269-1,059 cells/µL, NK cells: 136–880 cells/µL, B cells: 92–498 cells/µL.

IGRA test

Peripheral venous blood (3–4 ml) was collected from each participant and processed within six hours. Peripheral blood mononuclear cells (PBMCs) were isolated using Ficoll-Paque™ PLUS (GE Healthcare Life Sciences) density gradient centrifugation process. IGRA was performed using either an ELISPOT kit (Mingyuan, China) or an ELISA kit (Wantai Biological Pharmacy Enterprise, Beijing, China). This assay quantifies IFN-γ secretion from T cells following in vitro stimulation with Mtb-specific antigens (ELISPOT kit: CFP-10/ESAT-6 fusion protein; ELISA kit: CFP-10 and ESAT-6 recombinant protein), serving as an auxiliary diagnostic tool for Mtb infection.

Statistical analysis

Data analysis was conducted using GraphPad Prism version 10.0 (GraphPad Software, Inc., La Jolla, California, USA). Continuous variables were expressed as medians with interquartile ranges (IQRs). Between-group comparisons of nonparametric data were performed using the Mann-Whitney U test. Categorical variables were analyzed using the chi-squared test or Fisher′s exact test, as appropriate. Descriptive statistics summarized the baseline clinical characteristics of the study cohort. A two-tailed P-value ≤0.05 was considered statistically significant.

Results

Clinical characteristics

The study included 113 NTM-PD and 105 PTB patients. The baseline characteristics of both cohorts are summarized in Tables 1 and 2, and further illustrated in Fig. 2.

Table 1.

Clinical characteristics of the NTM-PD and PTB groups

Characteristics	NTM-PD group (n=113)	PTB group (n=105)	P value
Age (median, range)	61 (14-89)	52 (16-83)	<0.0001
Gender, n (%)			0.060
Male	50 (44.25%)	60 (57.14%)
Female	63 (55.75%)	45 (42.86%)
Smoking history, n (%)			0.001
Yes	20 (17.70%)	40 (38.10%)
No	93 (82.30%)	65 (61.90%)
Alcoholic history, n (%)			0.005
Yes	16 (14.16%)	32 (30.48%)
No	97 (85.84%)	73 (69.52%)
Symptoms, n (%)
Fever	25 (22.12%)	31 (29.52%)	0.219
Coughing and sputum	80 (70.80%)	63 (60.00%)	0.117
Hemoptysis	20 (17.70%)	7 (6.67%)	0.014
Chest tightness, Shortness of breath	39 (34.51%)	31 (29.52%)	0.470
Swollen lymph nodes	10 (8.85%)	23 (21.90%)	0.008

Values in bold indicate statistically significant differences (P < 0.05)

Abbreviations: NTM-PD non-tuberculous mycobacterial pulmonary disease, PTB pulmonary tuberculosis

Table 2.

Comparison of the clinical parameters between the NTM-PD and PTB groups

Characteristics	NTM-PD group (n=113)	PTB group(n=105)	P value
Blood routine examination
WBC (10^9/L)	6.40 (5.39, 8.01)	6.63 (5.04, 8.37)	0.925
RBC (10^9/L)	4.10 (3.80, 4.50)	4.47 (4.11, 4.76)	<0.001
Hb (g/L)	124 (111, 134)	132 (118, 142)	0.002
BPC (10^9/L)	243 (201, 283)	265 (211, 328)	0.031
ANC (10^9/L)	4.25 (3.28, 6.18)	4.36 (3.08, 5.74)	0.863
ALC (10^9/L)	1.34 (0.99, 1.90)	1.45 (1.08, 1.80)	0.474
AMC (10^9/L)	0.42 (0.34, 0.54)	0.43 (0.31, 0.55)	0.926
ESR (%)	30 (10, 61)	17 (7, 38)	0.029
Hs-CRP (mg/L)	4.29 (1.29, 34.49)	5.09 (1.60, 27.39)	0.657
Blood biochemical examination
PT (s)	11.6 (11.1, 12.1)	11.6 (11.2, 12.2)	0.476
APTT (s)	31.8 (29.6, 34.0)	32.9 (30.8, 35.2)	0.031
Alb (g/L)	38.4 (34.5, 42.2)	40.3 (37.5, 44.0)	0.024
PCT, n (%)			0.200
Positive	16 (21.33%)	10 (12.99%)
Negative	59 (78.67%)	67 (87.01%)
IGRA, n (%)			<0.0001
Positive	25 (24.51%)	82 (87.23%)
Negative	77 (75.49%)	12 (12.77%)
TB-Antibody
38KD, n (%)			0.099
Positive	63 (70.79%)	60 (82.19%)
Negative	26 (29.21%)	13 (17.81%)
38KD+16KD, n (%)			0.514
Positive	7 (7.87%)	3 (4.11%)
Negative	82 (92.13%)	70 (95.89%)
T lymphocyte subsets
CD3+T cell	1109 (793, 1468)	1083 (899, 1455)	0.999
CD4+T cell	602 (451, 866)	640 (526, 825)	0.642
CD8+T cell	417 (243, 568)	377 (316, 507)	0.836
NK cell	244 (148, 363)	177 (138, 323)	0.196
NKT cell	94 (39, 146)	84 (62, 107)	0.776
B cell	133 (81, 241)	169 (118, 258)	0.201
CD4+/CD8+	1.63 (1.02, 2.31)	1.67 (1.24, 2.05)	0.802

Values in bold indicate statistically significant differences (P < 0.05)

Abbreviations: NTM-PD non-tuberculous mycobacterial pulmonary disease, PTB pulmonary tuberculosis, PCT procalcitonin, IGRA interferon-gamma release assay

Fig. 2.

Disease profile and percentage of NTM-PD and PTB groups. Abbreviations: COPD, chronic obstructive pulmonary disease

Compared to the PTB group, the NTM-PD group was significantly older and exhibited lower rates of smoking and alcohol consumption. IGRA positivity was markedly less frequent in the NTM-PD group. Hemoptysis occurred significantly more frequently in the NTM-PD group, while lymph node enlargement was less common. No significant difference in gender distribution was observed between the two groups (Table 1).

Laboratory analysis revealed that NTM-PD patients had significantly lower levels of RBC, Hb, PC, and Alb, as well as a shorter APTT, compared to PTB patients. Conversely, ESR was significantly higher in NTM-PD group. No significant differences were observed in the remaining laboratory parameters (Table 2).

Regarding comorbidities, BE and COPD were significantly more prevalent in the NTM-PD group, whereas diabetes was more common in the PTB group (Fig. 2).

Characterization of the BALF microbiome in the NTM-PD and PTB groups

We compared the microbiota of the lower respiratory tract between groups using BALF culture and tNGS, revealing distinct microbial signatures (see Fig. 3 and Supplementary Table S1). The tNGS analysis showed that the number of sequencing reads across all samples ranged from 43 to 55,630, with a median of 426 reads.

Fig. 3.

Comparing the proportions of the microbiome detected in the NTM-PD and PTB groups. A G⁻ bacteria. B G⁺ bacteria. C Fungi

Among Gram-negative (G⁻) bacteria, Pseudomonas aeruginosa was more frequently detected in the NTM-PD group (P < 0.05), while Neisseria spp. was more prevalent in the PTB group (P < 0.05) (Fig. 3A). Among Gram-positive (G⁺) bacteria, Streptococcus viridans was less prevalent in the NTM-PD group (P < 0.05) (Fig. 3B). Aspergillus spp. was significantly more prevalent in the NTM-PD group (Fig. 3C).

Notably, Nocardia spp. were present in 4.42% of NTM-PD patients, but were absent in the PTB group (see Fig. 3B). On the contrary, Penicillium spp. were only detected in the PTB group (1.90%), with no cases identified in NTM-PD patients (Fig. 3C). Although these differences did not reach statistical significance, likely due to the limited sample size, it is a clinically relevant finding.

Associations between the microbiome and clinical characteristics and immune function in patients with NTM-PD

Microbiome detection patterns in NTM-PD patients varied with host characteristics (Fig. 4 and Supplementary Table S2). As shown in Fig. 4A, Klebsiella pneumoniae was significantly more frequent in male patients (P < 0.05). Candida spp. showed a non-significant trend towards higher detection in males. Neisseria spp. was more prevalent in patients aged ≤60 years, though this was not statistically significant. In PCT-negative patients, Neisseria spp. and Streptococcus viridans were more frequent, while Candida spp. was less common. No significant differences in microbiota were observed based on IGRA results or BE presence.

Fig. 4.

Microbiome detection proportions in NTM-PD patients stratified by (A) clinical factors and (B) immune status. Reference values: CD3⁺≥834 cells/µL, CD4⁺≥395 cells/µL, CD8⁺≥269 cells/µL, B≥92 cells/µL, and NK≥ 136 cells/µL [24]. Abbreviations: IGRA, interferon-gamma release assay; PCT, procalcitonin; BE, bronchiectasis

Figure 4 B illustrates the correlations between the microbiome and the host’s immune function. The detection rate of Neisseria spp. was significantly lower in patients with low absolute counts of CD3⁺ and CD4⁺ T cells (CD3⁺ P > 0.05, CD4⁺ P < 0.05). Conversely, Pseudomonas aeruginosa and Candida spp. were more frequently detected in patients with reduced CD4⁺ T cell counts (P < 0.05). No significant differences in microbial detection were observed based on CD8⁺ T, NK, or B cell counts. Notably, all four NTM-PD patients with Klebsiella pneumoniae had normal lymphocyte subset counts, though this was not statistically significant due to the small sample size.

Microbiome distribution across different NTM species

Among the 113 NTM-PD patients, the overall NTM detection rate was 82.30% (93/113). The remaining 20 patients with negative NTM cultures were diagnosed based on positive molecular detection (NTM nucleic acid testing) of sputum and/or BALF samples. Notably, four patients (4.42%, 4/93) were co-infected with two distinct NTM species, which resulted in a discrepancy between the total number of species identifications (n = 97) and the actual number of patients (n = 93). Specifically, three patients were co-infected with M. abscessus and M. intracellulare, and one patient was co-infected with M. abscessus and M. avium.

The most prevalent species identified were the M. avium complex (MAC) (62.37%, 58/93), followed by the M. abscessus complex (MABC) (27.96%, 26/93), and M. kansasii (6.45%, 6/93). Within MAC, M. intracellulare predominated (48.28%, 28/58), followed by M. avium (20.69%, 12/58), M. maeseillense (3.45%, 2/58), and M. avium-intracellulare complex (27.58%, 16/58). Within MABC, 80.77% (21/26) were M. abscessus, whereas M. chelonae (11.54%, 3/26) and M. chelonae-abscessus complex (7.69%, 2/26) were less prevalent. Other NTM species accounted for a further 7.53% (7/93) of isolates, including M. scrofulaceum, M. gordonae, M. fortuitum, M. shinjukuense, M. iranicum and M. nonchromogenicum. Microbiome distribution across NTM species is shown in Fig. 5, Supplementary Table S2 and S3.

Fig. 5.

Distribution of microbiome detected in patients infected with different NTM species. Abbreviations: MAC, M. avium complex. MABC, M. abscessus complex

Comparative analysis of the microbiome across different NTM species groups revealed several trends, although no statistically significant differences were observed. Pseudomonas aeruginosa and Staphylococcus spp. were detected more frequently in patients infected with MAC (17.24% and 12.07%, respectively) and MABC (19.23% and 23.08%, respectively), in contrast to their absence in the M. kansasii and ‘Others’ groups. Klebsiella pneumoniae was present at a higher proportion in the M. kansasii group (16.67%), but absent in the ‘Others’ group. Aspergillus spp. detection was similar across MAC, MABC, and M. kansasii but higher among patients with less common NTM species. These patterns suggest potential species-specific airway microbial environments, warranting further investigation in larger studies.

Discussion

This study systematically compared the clinical characteristics, laboratory parameters, immune status and BALF microbiome composition between patients with NTM-PD and PTB. We further explored microbiome differences among patients infected with various NTM species. Our findings revealed distinct host characteristics, immune profiles, and lung microbiome signatures between the two diseases, providing multifaceted insights into the unique susceptibility and pathogenesis of NTM-PD. These results could inform improved differential diagnosis and understanding the pathogenesis of NTM-PD.

Consistent with prior studies [4, 25, 26], NTM-PD patients were significantly older and had lower rates of smoking and alcohol consumption than PTB patients. NTM primarily acts as an opportunistic pathogen affecting older individuals with impaired mucociliary clearance or chronic lung disease, whereas lifestyle-related risk factors such as tobacco and alcohol use are more strongly associated with susceptibility to Mtb infection. Hemoptysis was more prevalent in NTM-PD patients, whereas lymphadenopathy was more frequent in PTB patients, aligning with previous reports [27, 28]. This likely reflects differences in pathogenesis: Mtb is phagocytosed by macrophages and transported to regional lymph nodes, triggering a robust immune response and characteristic lymphadenopathy, whereas NTM typically causes localized airway infection with less lymphatic dissemination [29]. Furthermore, IGRA positivity was significantly lower in NTM-PD patients than in PTB patients, a finding corroborated by multiple studies [4, 30]. This discrepancy arises because the IGRA antigens ESAT-6 and CFP-10 are encoded by genes located in the RD1 region, which is present in the M. tuberculosis complex, but absent in most NTM species [31]. Consequently, T cells from NTM-PD patients are generally not activated by these antigens, resulting in negative IGRA outcomes. However, certain NTM species, such as M. kansasii and M. gordonae, possess the RD1 region or homologous genes and may therefore cause positive IGRA results [31–33]. Among 25 IGRA-positive NTM-PD patients in this study, only four were infected with RD1-containing M. kansasii or M. gordonae, the remaining positive cases were likely due to old TB lesions. All patients were diagnosed based on positive NTM culture and/or molecular testing with concurrent negative Mtb testing. This highlights that IGRA cannot reliably distinguish active PTB from NTM-PD in individuals with prior TB exposure. Additionally, NTM-PD was also more frequently associated with underlying BE and COPD, while PTB was more often complicated by diabetes, consistent with prior reports [5, 34]. Structural lung damage caused by BE or COPD is known to increases susceptibility to NTM infection, whereas diabetes-induced immune dysfunction may promote the progression to active TB.

NTM-PD patients exhibited lower levels of RBC, Hb, PC, and Alb, along with elevated ESR, indicating a state of chronic inflammation, poorer nutritional status, and inflammatory-metabolic consumption [35–38]. These findings likely reflect the prolonged course, suboptimal immune status, and sustained microbial burden characteristics of NTM-PD.

A key finding of this study emerged from the parallel analysis of the BALF microbiome using tNGS and culture. This analysis demonstrated that patients with NTM-PD harbor a more diverse lung microbiome, with a greater propensity for co-colonization or co-infection with conditional pathogens and fungi. NTM-PD patients showed a higher detection rates of Pseudomonas aeruginosa, Aspergillus spp., and Nocardia spp., whereas Streptococcus viridans was more frequently in PTB patients. This suggests that NTM-PD patients are more susceptible to pulmonary aspergillosis or severe Pseudomonas aeruginosa infection. This phenomenon is likely closely linked to impaired host defense mechanisms and alterations in the local microenvironment. NTM-PD patients often have extensive and persistent structural lung diseases, such as BE, cavity formation or cystic fibrosis [39], which lead to airway structural damage, mucus retention, impaired mucociliary clearance and localized hypoxia. These abnormalities create favorable conditions for the colonization, germination and biofilm formation of Aspergillus spp. conidia and Pseudomonas aeruginosa [40–42]. Moreover, NTM infection may alter local immune responses (e.g. macrophage dysfunction and enhanced Th2 inflammation) [43], further hindering clearance of Aspergillus spp. and Pseudomonas aeruginosa. In contrast, the robust Th1 response during the acute inflammatory phase in PTB may limit early fungal colonization, and cavities often resolve with effective therapy, reducing persistent colonization opportunities [44, 45]. Colonization or infection with Aspergillus spp. in NTM-PD patients can lead to the formation of aspergilloma, causing physical obstruction and irritation. Through mechanisms such as amplified inflammatory responses and exacerbated tissue damage, it can also significantly increase the risk of hemoptysis, make treatment more challenging and induce allergies. This can markedly worsen respiratory symptoms (cough, sputum, dyspnea, wheezing and hemoptysis) and systemic symptoms (fever, fatigue and weight loss) [46–49]. This represents not merely a superimposed infection, but rather a complex, mutually reinforcing vicious cycle with NTM infection. This leads to faster disease progression, more severe symptoms, greater treatment difficulty, and a poorer prognosis. Similarly, colonization or infection with Pseudomonas aeruginosa significantly exacerbates airway inflammation, increases mucus secretion, further impairs ciliary function and accelerates lung tissue destruction and functional decline. This contributes to a protracted and severe clinical course in NTM-PD patients [50]. Furthermore, Pseudomonas aeruginosa often exhibits strong antibiotic resistance and readily forms biofilms. Its virulence factors and persistent inflammatory stimulation may also interfere with the host’s immune clearance of NTM [51]. Therefore, the presence of Aspergillus spp. or Pseudomonas aeruginosa considerably complicates the treatment of NTM-PD. Additionally, the detection of Nocardia spp. in NTM-PD patients is an emerging risk factor, suggesting that immune compromise or chronic airway inflammation may facilitate the entry of exogenous or opportunistic pathogens [52, 53]. In contrast, the higher prevalence of Neisseria spp. in PTB patients may be related to the migration of upper respiratory tract flora, the location of airway lesions or the characteristics of lesions in PTB. Taken together, these findings suggest that NTM-PD is a complex disease driven by interactions between the pathogen, host and microbiome rather than by a single pathogen infection.

An unexpected finding was that Penicillium spp. were detected exclusively in the BALF of PTB patients, despite all subjects being HIV-negative. Previous studies have shown that co-infection with Penicillium and Mtb is predominantly associated with HIV-related immunosuppression [54, 55], highlighting the susceptibility of severely immunocompromised hosts to Penicillium infection. However, colonization by Penicillium spp. may also occur in HIV-negative PTB patients, potentially attributable to local airway damage, microbial dysbiosis, or immune dysfunction induced by active TB. Nevertheless, given the relatively low detection rate and the cross-sectional design of this study, these findings should be interpreted cautiously and validated in larger cohorts. Notably, the complete absence of Penicillium spp. in NTM-PD patients, despite their older age and frequent presented with underlying structural lung disease, further suggests that PTB and NTM-PD may harbor distinct lower airway microecological niches.

Within the NTM-PD cohort, detection rates of Aspergillus spp. and Pseudomonas aeruginosa tended to increase with age, particularly among patients aged ≥ 60 years. This finding is consistent with prior studies [56–58] regarding the increased risk of Aspergillus spp. or Pseudomonas aeruginosa infection in elderly patients with lung diseases who reportedly experience poorer outcomes. Notably, NTM-PD patients with normal absolute NK cell and B lymphocyte counts exhibited a higher Aspergillus spp. detection rate than those with subnormal counts. This may reflect the proliferation of NK cells in response to persistent infection in NTM-PD patients co-colonized with Aspergillus spp., indicating an ongoing and challenging immune response. Concurrently, Aspergillus spp. can promote allergy and airway remodeling though Th2 responses while simultaneously suppressing anti-NTM Th1 immunity. This mechanism likely contributes to treatment resistance and may explain the higher detection rate among patients with normal B cell counts [59]. In contrast, Pseudomonas aeruginosa was more frequently detected in NTM-PD patients with subnormal absolute counts of CD3⁺, CD4⁺ T cells, and B lymphocytes. This pattern is consistent with the known predilection of Pseudomonas aeruginosa for infecting immunocompromised hosts.

Candida spp. are common commensal fungi that often colonize the oral cavity or respiratory tract. While typically non-pathogenic, they can become opportunistic pathogens in immunocompromised hosts. In this study, Candida spp. was detected in both groups, indicating their widespread colonization. Interestingly, a slightly higher rate of Candida spp. isolation from the lower respiratory tract was observed in male compared to female NTM-PD patients. This finding is supported by Zhou et al. [60] and may be attributable to higher smoking rates, an increased prevalence of chronic lung diseases or differential immune responses in males. Furthermore, NTM-PD patients with subnormal absolute CD4⁺ T cell counts exhibited a significantly higher Candida spp. detection rate than those with normal counts, indicating that immunocompromised individuals with NTM-PD are more susceptible to Candida spp. invasion. This susceptibility stems from the central role of Th1 immunity and IFN-γactivity in systemic antifungal defense, supported by Th17-mediated mucosal protection [61, 62]. Impairment of these immune pathways substantially increases the risk of invasive Candida spp. This observation underscores the critical role of immune status in driving microbiome variations and influencing disease progression in NTM-PD, particularly in individuals with low CD4⁺ T cell counts. In addition, another study suggested that PCT levels may be elevated in patients with lower respiratory tract fungal infections, with the degree of elevation often correlating with the severity of invasive disease [63]. This aligns with our finding that the Candida spp. detection rate was markedly higher in PCT-positive NTM-PD patients, suggesting that PCT positivity in NTM-PD may indicate more extensive fungal invasion.

Neisseria spp. and Streptococcus viridans were commonly isolated in BALF cultures of both NTM-PD and PTB patients. The observed differences in detection rates between the two groups may reflect distinct pathological mechanisms, host immune responses, and microecological niches associated with NTM versus Mtb infection. We propose the following hypothetical explanations: (1) NTM-PD patients often have structural lesions such as BE or cystic fibrosis, which can lead to impaired mucociliary clearance and reduced microbial diversity (and thus fewer commensals such as Neisseria spp. and Streptococcus viridans) [64]; (2) NTM-PD often presents as an opportunistic infection characterized pathologically by granulomas and can produce scant and viscous sputum, which lacks the favorable substrate for the proliferation of these bacteria; and (3) the high virulence of Mtb can induce a robust Th1 immune response, resulting in caseous necrosis and cavity formation in the lungs. This intense inflammatory environment may disrupt the respiratory mucosal barrier, potential facilitating the overgrowth or colonization of these bacteria [65].

Although statistical significance was not reached—likely due to the limited sample size—certain trends in microbiome composition were observed across different NTM species. Notably, the detection rates of Pseudomonas aeruginosa and Staphylococcus spp. were higher in patients infected with MAC or MABC than in those infected with M. kansasii. This observation is indirectly supported by the findings of Luis et al. [66], who reported that cells co-infected with MAC and Mtb promoted the growth of Pseudomonas aeruginosa, whereas those infected with M. kansasii did not. Therefore, we hypothesize that MAC and MABC, which frequently colonize the lungs and are particularly associated with structural lung damage such as BE and cavitation, disrupt the respiratory mucosal barrier and impair secretion clearance. This creates favorable conditions for the colonization and proliferation of Pseudomonas aeruginosa (which prefers moist, hypoxic conditions) and Staphylococcus spp. (which readily colonize damaged mucosa). In contrast, M. kansasii infection is less frequently linked to severe BE or extensive structural destruction, potentially preserving mucosal integrity and thereby resulting in lower detection rates of these co-colonizing bacteria.

It should be explicitly acknowledged that the cross-sectional design of this study cannot draw definitive conclusions regarding the causal relationship between microbiome alterations and disease status. While we interpret the distinct microbiome profiles observed in NTM-PD patients as likely consequences of underlying structural lung damage and immune dysregulation, we cannot exclude the possibility that specific microbial colonization patterns may predispose individuals to NTM infection or TB progression. For instance, pre-existing dysbiosis characterized by the dominance of Pseudomonas aeruginosa could theoretically impair mucosal immune defense, thereby facilitating NTM colonization [50, 67, 68].

Furthermore, significant differences in baseline characteristics between the two groups, including age, smoking history, alcohol consumption, and comorbidities (such as BE, COPD, diabetes), may independently influence the composition of the respiratory microbiome. In particular, smoking can directly alter airway microbial ecology by impairing mucociliary clearance and promoting the enrichment of pro-inflammatory bacterial species [69, 70]. Additionally, the higher prevalence of structural lung diseases (BE/COPD) in NTM-PD patients may create distinct ecological niches that favor the colonization of opportunistic pathogens regardless of NTM infection status [20, 71]. Similarly, diabetes-associated immune dysfunction in PTB patients may selectively alter or suppress certain commensal microbial communities [72]. Therefore, longitudinal cohort studies tracking microbiome dynamics from pre-disease stages through to disease progression are required to understand the complex interplay among host factors, changes to the microbial community, and disease pathogenesis.

Despite identifying unique microbial community features in NTM-PD patients compared to PTB patients, this study has several limitations: (1) Sample size constraints. Although over 100 NTM-PD patients with complete microbiological data were enrolled over five years — a considerable number for this relatively rare disease —the sample size remains insufficient for robust statistical analysis of associations between specific microbial species and immune parameters, limiting the ability to detect significant correlations. (2) Single-center retrospective design. This design may introduce selection bias, meaning the results may reflect characteristics specific to our local population. Microbial communities are influenced by various factors, including comorbid conditions, environmental exposures and lifestyle, all of which contribute to high inter-individual variability. Future multicenter studies are needed to validate and generalize these findings. (3) Heterogeneity of NTM species. NTM encompasses numerous species with varying pathogenicity, clinical courses, and treatment regimens. Aggregating all NTM species into a single ‘NTM-PD’ category may obscure species-specific microbiome associations. However, even after subdividing by NTM species, the detection rates for specific microbiota were too low for detailed analysis. Future research should focus on conducting multicenter prospective cohorts to enhance representativeness, integrating multi-omics approaches (tNGS + metagenomics + metabolomics) to explore the ‘microecological signature’ of NTM-PD, investigating the mechanisms underlying microbiome-immune-structural lung disease interactions, and developing diagnostic or predictive models based on microbiome and immune markers. (4) Causal inference. Due to the cross-sectional design of this study, it is not possible to establish temporal and causal relationships between microbiome composition and disease development. The observed associations may reflect one of the following: (a) microbiome changes resulting from disease-related tissue damage and immune system dysfunction; (b) pre-existing dysbiosis that makes individuals more susceptible to pathogen colonization and disease progression; or (c) the impact of unmeasured host or environmental factors. Additionally, significant baseline differences between the two groups (including age, smoking history, and comorbidities) may independently influence respiratory microbiome composition, thereby further complicate causal interpretation.

Conclusion

In conclusion, this study reveals significant differences in clinical characteristics, immune function, and airway microbiome composition between NTM-PD and PTB. NTM-PD was more strongly associated with advanced age, structural lung disease, nutritional decline and chronic inflammation. The airway microbiome of NTM-PD patients was characterized by higher rates of opportunistic bacterial and fungal colonization, which are strongly influenced by the host immune status, especially CD4⁺ T cell count. Trends observed at the species level among NTM subgroups suggest the existence of distinct host–microbiome interactions. These findings highlight NTM-PD as a complex disease driven by the interplay of host susceptibility, immunological impairment, and microbial dysbiosis. An integrated assessment of microbiome and immune profiles could enhance discrimination between NTM-PD and PTB, help to identify high-risk co-infections and support the development of personalized management strategies. Future multicenter prospective studies incorporating multi-omics approaches are needed to elucidate the mechanistic pathways governing host–microbe interactions in NTM-PD and to guide precision diagnosis and therapy.

Abbreviations

NTM

Non-tuberculous mycobacteria

NTM-PD

Non-tuberculous mycobacterial pulmonary disease

PTB

Pulmonary tuberculosis

Mtb

Mycobacterium tuberculosis

BE

Bronchiectasis

BALF

Bronchoalveolar lavage fluid

mNGS

Metagenomic next-generation sequencing

tNGS

Targeted next-generation sequencing

IGRA

Interferon-gamma release assay

PNB/TCH

Paranitrobenzoic acid/2-thiophenecarboxylic acid hydrazide

WBC

White blood cell count

RBC

Red blood cell count

Hb

Hemoglobin

PC

Platelet count

ANC

Absolute neutrophil count

ALC

Absolute lymphocyte count

AMC

Absolute monocyte count

ESR

Erythrocyte sedimentation rate

hs-CRP

High-sensitivity C-reactive protein

PT

Prothrombin time

APTT

Activated partial thromboplastin time

Alb

Albumin

PCT

Procalcitonin

COPD

Chronic obstructive pulmonary disease

PBMCs

Peripheral blood mononuclear cells

G⁻

Gram-negative

G⁺

Gram-positive

MAC

M. avium complex

MABC

M. abscessus complex

Authors’ contributions

JM contributed to conceptualization, data curation, writing original draft, writing-review & editing, formal analysis. YD contributed to software, investigation, formal analysis. ZYX contributed to data curation, investigation. YL contributed to supervision, resources, methodology, conceptualization. XQW contributed to writing-review & editing, writing-original draft, visualization, supervision, resources, project administration. JQL contributed to writing-review & editing, funding acquisition, formal analysis, conceptualization.

Funding

This project was financially supported by the Capital′s Funds for Health Improvement and Research (Grant No. 2024-2-5094).

Data availability

All data supporting the findings of this study are available within the paper and its Supplementary Information.

Declarations

Ethics approval and consent to participate

This retrospective study was conducted among hospitalized patients and was approved by the Ethics Committee of the 8th Medical Center of the PLA General Hospital (approval number: 309202109061503). This was a retrospective study that involved no patient identifiers; therefore, the Ethics Committee of the 8th Medical Center of PLA General Hospital waived the requirement for written informed consent. However, our study complies with Helsinki′s Ethical Principles for Medical Research.

Consent for publication

Not Applicable.

Competing interests

The authors declare no competing interests.