A new strategy for the diagnosis of tuberculosis based on secreted antigens: evaluation of the efficacy of alveolar lavage ESAT-6 and CFP-10 tests

Abstract

Objectives

To assess the diagnostic value of Mycobacterium tuberculosis (MTB)-specific secreted antigens ESAT-6 and CFP-10 in bronchoalveolar lavage fluid (BALF), and to explore novel adjunctive diagnostic strategies to enhance the diagnostic rate of pulmonary tuberculosis (PTB).

Methods

104 patients with PTB (59 confirmed cases and 45 clinically diagnosed cases) and 72 patients with non-tuberculosis lung diseases (control group), hospitalized from May 2021 to July 2023, underwent bronchoscopy. The concentrations of ESAT-6 and CFP-10 antigens were detected by enzyme-linked immunosorbent assay (ELISA). Optimal cutoff values were determined by receiver operating characteristic (ROC) curves to evaluate antigen diagnostic capability for active PTB, compared with acid-fast bacilli (AFB) and Xpert MTB/RIF.

Results

Antigen expression characteristics: Concentrations of ESAT-6 and CFP-10 were significantly higher in the PTB group compared to controls (both P < 0.001). ESAT-6 levels did not differ significantly between confirmed and clinically diagnosed cases (P > 0.05), whereas CFP-10 concentrations were significantly higher in confirmed cases (P = 0.045). Comparison of diagnostic efficacy: Sensitivities of AFB and Xpert MTB/RIF were 26.92% (95% CI 18.6–37.2%) and 56.73% (95% CI 46.1–66.8%). ESAT-6 and CFP-10 demonstrated significantly higher sensitivities (77.89%, 95% CI 68.5–85.1%; 67.31%, 95% CI 57.3–76.0%) compared to AFB (Δ = 50.97%, P < 0.001; Δ = 40.39%, P < 0.001) and Xpert MTB/RIF (Δ = 21.16%, P < 0.001; Δ = 10.58%, P = 0.021), but had lower specificity (P < 0.001). Combined testing strategy: Parallel testing (either antigen positive) yielded sensitivity of 94.23% (95% CI 87.4–97.6%) and negative predictive value of 99.2%, while tandem testing (both antigens positive) provided specificity of 95.83% (95% CI 88.1–98.6%). Subgroup analysis: No statistically significant difference was observed in antigen sensitivities between bacteriologically positive and negative PTB groups (P > 0.05).

Conclusions

(1) ESAT-6 and CFP-10 detection in BALF significantly improves PTB diagnostic sensitivity, unaffected by bacterial load, particularly benefiting the diagnosis of bacteriologically negative PTB. (2) Combined antigen testing strategies (parallel/tandem) optimally balance sensitivity and specificity, meeting clinical requirements for ruling out or confirming PTB diagnosis.

Introduction

Tuberculosis (PTB) is a chronic infectious disease caused by Mycobacterium tuberculosis infection and has surpassed AIDS as the world's leading cause of death from infectious diseases. China is among the countries with the highest PTB burden. According to the World Health Organization's (WHO) Global Tuberculosis Report 2024 [1], there were approximately 10.8 million new PTB cases globally in 2023, with China accounting for 741,000 cases, ranking third among the 30 high-burden countries. In 2014, WHO initiated the End TB Strategy, aiming to reduce TB incidence to less than 10 per 100,000 population by 2035 [2]. A core component of this strategy is "early and accurate diagnosis," crucial for breaking the transmission chain. Epidemiological data indicate that over 60% of global PTB cases are initially diagnosed in primary healthcare centers (PHCs) [1], highlighting the importance of PHCs in PTB control. However, current primary diagnostic methods, primarily sputum smear microscopy, suffer from low sensitivity, missing 35–40% of cases [3], thus exacerbating delays in diagnosis and increasing risks for secondary drug resistance. Therefore, there is an urgent need to identify highly sensitive, cost-effective, and user-friendly diagnostic tools to bridge this diagnostic gap and facilitate the implementation of the End TB Strategy.

Currently, tuberculosis diagnosis relies on identifying etiological evidence, primarily through MTB acid-fast bacilli (AFB) smears or mycobacterial cultures. AFB smear microscopy is economical and efficient but exhibits low sensitivity, particularly in patients with low bacterial loads, and cannot differentiate between non-tuberculous mycobacteria (NTM) and MTB [3, 4]. Mycobacterial culture, considered the gold standard, offers superior sensitivity but faces significant practical and technical limitations, including prolonged incubation periods and stringent biosafety requirements, delaying timely diagnosis and treatment initiation [5, 6]. Molecular diagnostic techniques, notably polymerase chain reaction (PCR), have significantly improved PTB diagnosis due to their high sensitivity, specificity, rapidity, and cost-effectiveness [7]. Advanced molecular diagnostics such as Xpert MTB/RIF and Xpert MTB/RIF Ultra have further enhanced MTB detection, particularly in HIV-positive patients, and enabled rapid identification of drug-resistant strains [8]. Nevertheless, the high costs and strict operational requirements limit their widespread adoption in primary care settings [9].

Immunological tests, detecting host immune responses to MTB-specific antigens, also play a significant auxiliary role in PTB diagnosis. Gamma-interferon release assays (IGRAs) and the tuberculin skin test (TST) are commonly utilized for detecting latent MTB infection (LTBI) by assessing T-cell responses to MTB-specific antigens ESAT-6 and CFP-10 [10]. However, these tests have limitations in distinguishing active PTB from LTBI, influenced by factors, such as the host's immune status and prior infection history [11, 12]. Given the high diagnostic potential of ESAT-6 and CFP-10—specifically attributable to their secretion by actively replicating M. tuberculosis—this study leverages this property by directly quantifying these antigens in alveolar lavage fluid via ELISA. Unlike methods that measure the host immune response, our approach targets the antigens themselves as direct markers of active pulmonary tuberculosis. We systematically compared the sensitivity and specificity of this ELISA-based antigen detection against traditional AFB smear microscopy and Xpert MTB/RIF assays to evaluate its diagnostic utility and propose a novel clinical strategy for PTB diagnosis.

Materials and methods

Study design and participants

This was a retrospective, single-center clinical study conducted at the Affiliated Hospital of Chuanbei Medical College between May 2021 and July 2023. The study included patients who underwent fiberoptic bronchoscopy. Clinical history data and bronchoalveolar lavage fluid samples were systematically collected. Patients were classified into pulmonary tuberculosis (PTB) and non-pulmonary tuberculosis (non-PTB) groups based on diagnoses. The PTB group was further subdivided into bacteriologically confirmed cases of pulmonary tuberculosis (BC-PTB) and clinically diagnosed cases of pulmonary tuberculosis (CD-PTB). Patients in the PTB group met the following conditions: (1) fulfill the diagnostic criteria outlined in WS288-2017 [13] and (2) exhibit effectiveness after standardized anti-tuberculosis treatment. To ensure accurate microbiological test results, BALF samples from all PTB patients were collected before initiating anti-tuberculosis therapy. Inclusion criteria for the non-PTB group were: (1) no history or exposure to tuberculosis; (2) definitive diagnosis of pulmonary tumors by pathological examination or other pulmonary infections with confirmed etiological evidence; and (3) negative IGRAs results. In addition, all study participants were tested to exclude hepatitis C virus, hepatitis B virus, and HIV infections.

Collection and preservation of BALF

After obtaining informed consent, BALF collection strictly adhered to technical guidelines provided by the Chinese Thoracic Society [14] and standard operating procedures. Alveolar lavage involved three lavages using sterile saline, each with a volume of 20–60 mL, totaling 100–300 mL. Fluid recovery was conducted at a negative pressure of − 3.3 to − 13.3 kPa after each lavage, achieving a recovery rate of 40–70%. The collected fluid was immediately centrifuged at [500] g for [15] min to pellet cellular debris. The cell-free supernatant was then carefully aliquoted into sterile frozen tubes pretreated with silicone oil, and immediately stored at − 20 °C. To prevent sample degradation, each sample underwent a single freeze–thaw cycle, with thawing performed overnight at 4 °C before use.

Optimal sample dilution identification

Checkerboard titration was performed to determine optimal sample dilution. Three samples with the highest acid-fast bacilli content (smear results ranging from +++ to ++++) were pre-tested at dilution ratios of 1:1, 1:5, 1:10, and 1:20. The optimal dilution was identified by calculating the ratio (P/N value) of absorbance between positive samples and negative controls. Experimental results indicated that undiluted samples provided the highest P/N value and thus were used in subsequent assays.

ELISA detection of BALF

BALF supernatant was analyzed using a Mycobacterium tuberculosis antigen ELISA kit (Camilo Biological Co., Nanjing, China) in strict accordance with the manufacturer’s instructions. Briefly, the kit was equilibrated to room temperature for 20 min before use. Lyophilized ESAT-6 and CFP-10 standards were reconstituted in 1 mL of standard diluent to yield stock concentrations of 100 ng/mL (ESAT-6) and 20 ng/mL (CFP-10). Serial twofold dilutions (100–1.5%) were prepared to construct the standard curves. Standards and BALF samples (100 µL/well, in duplicate) were loaded onto antibody-coated microplates and incubated at 37 °C for 90 min. After two washes, 100 µL of 1:100-diluted biotinylated detection antibody was added, followed by incubation at 37 °C for 60 min. Plates were washed three times, and 100 µL of horseradish–peroxidase conjugate was applied for 30 min at 37 °C in the dark. Following five washes, 100 µL of substrate solution was added and the reaction was stopped with 2 M H₂SO₄ when adequate color developed. The concentrations of ESAT-6 and CFP-10 in BALF samples were interpolated from a standard curve generated using recombinant protein standards, following the manufacturer's protocol. The standard curve was plotted (OD value vs. concentration) and fit using a four-parameter logistic (4PL) model to convert optical density (OD) readings into quantitative concentrations (ng/mL).

Clinical data collection in patients with and without PTB

For PTB participants, demographic and clinical data—including age, sex, ethnicity, comorbidities, HIV status, chest CT findings (presence or absence of cavities), prior PTB history, and response to anti-tuberculosis therapy—were recorded. For non-PTB participants, age and sex were collected.

Statistical analysis

Statistical analyses were performed using SPSS version 25 (IBM Corp.). The optimal diagnostic cutoff values for ESAT-6 and CFP-10 were determined by receiver operating characteristic (ROC) curve analysis. The point on the curve that maximized Youden's Index (J = sensitivity + specificity − 1) was selected as the optimal threshold. Data normality was evaluated using the Shapiro–Wilk test. Normally distributed variables were presented as mean ± standard deviation and compared with the independent-samples t test. Non-normally distributed data were expressed as median (interquartile range) and compared using the Mann–Whitney U test. Categorical variables were analyzed with Pearson’s χ² test. A two-sided P value < 0.05 was considered statistically significant. GraphPad Prism 8.0.1 was used for data visualization.

Results

Basic characteristics of study participants

Between May 2021 and July 2023, 104 patients with PTB (59 bacteriologically confirmed [BC-PTB] and 45 clinically diagnosed [CD-PTB]) underwent bronchoscopy and were enrolled. All PTB cases responded to standard anti-tuberculosis therapy. The control cohort comprised 72 patients with non-tuberculous pulmonary diseases confirmed by microbiological testing. Baseline characteristics are summarized in Table 1. Participant ages ranged from 13 to 85 years (median 50 years), including both adult and pediatric patients. In addition, 75.3% were male. Cough was the most common symptom (69.3%), and diabetes mellitus was the most frequent PTB comorbidity (11.5%).

Table 1.

Baseline demographic and microbiological characteristics of the study participants (n = 176)

	PTB group (n = 104)	Non-PTB group (n = 72)	p value
Median age [IQR] years	44 [13–78]	58.5 [16–85]	0.0001*
Men (%)	75 (72.1%)	48 (66.7%)	0.370***
Previous pulmonary tuberculosis (%)	1 (0.96%)	0 (0%)	1.00**
Cough (%)	68 (65.3%)	54 (75.0%)	0.174***
Fever (%)	14 (13.4%)	23 (31.9%)	0.003***
Weight loss (%)	19 (18.2%)	11 (15.2%)	0.604***
Night sweats (%)	27 (25.9%)	5 (6.9%)	0.001***
Diabetes mellitus (%)	12 (11.5%)	9 (12.5%)	0.843***
Hypertension (%)	2 (1.9%)	9 (12.5%)	0.018**
Extrapulmonary tuberculosis (%)	7 (6.7%)	0 (0%)	0.045**
Lung tumour (%)	3 (2.9%)	8 (11.1%)	0.045**
Sepsis (%)	1 (1.0%)	1 (1.4%)	1.00**
Connective tissue disease (%)	0 (0%)	5 (6.9%)	0.018**

*Mann–Whitney U test

**Fisher's exact test

***Chi-square test

BALF ESAT-6 and CFP-10 concentrations

Median BALF concentrations of ESAT-6 and CFP-10 were significantly higher in both BC-PTB and CD-PTB groups than in the non-PTB group (P < 0.001; Fig. 1). ESAT-6 concentrations did not differ between BC-PTB and CD-PTB (P = 0.056), whereas CFP-10 levels were higher in BC-PTB (P = 0.045). ROC analysis yielded areas under the curve of 0.856 for ESAT-6 and 0.805 for CFP-10. Optimal cutoff values were 24.54 ng/mL for ESAT-6 and 4.895 ng/mL for CFP-10 (Fig. 2).

Figure 1.

Box-plot distributions of BALF ESAT-6 and CFP-10 concentrations in CD-PTB, BC-PTB, and non-PTB groups (n = 176)

Figure 2.

Receiver-operating-characteristic curves for ESAT-6 and CFP-10 in the diagnosis of pulmonary tuberculosis (n = 176)

Diagnostic performance of ESAT-6 and CFP-10

Using these cutoffs, ESAT-6 achieved a sensitivity of 77.9% and a specificity of 79.2%, while CFP-10 achieved 67.3% and 77.8%, respectively (Table 2; Fig. 3). Both antigens were significantly more sensitive than acid-fast bacilli (AFB) smear (26.9%) and Xpert MTB/RIF (56.7%), albeit with reduced specificity (P < 0.001 for ESAT-6; P < 0.05 for CFP-10). Positive and negative predictive values were 84.4% and 71.3% for ESAT-6 and 81.4% and 62.2% for CFP-10.

Table 2.

Diagnostic performance of ESAT-6, CFP-10, sputum acid-fast bacillus smear, and Xpert MTB/RIF for active pulmonary tuberculosis (n = 176)

	PTB (n = 104)	non-PTB (n = 72)	Sensitivity (%)	Specificity (%)	PPV (%)	NPV (%)
AFB
+	28	0	26.92 (18.69–36.51)	100.00 (95.01–100)	100.00 (87.66–100)	48.65 (40.36–56.99)
−	76	72
GeneXpert MTB/RIF
+	59	0	56.73 (46.65–66.41)	100.00 (95.01–100)	100.00 (93.94–100)	61.54 (52.09–70.38)
−	45	72
ESAT-6
+	81	15	77.89 (68.47–85.19)	79.18 (67.67–87.50)	84.38 (75.22–90.70)	71.25 (59.88–80.55)
−	23	57
CFP-10
+	70	16	67.31 (57.32–76.00)	77.78 (66.15–86.39)	81.39 (71.25–88.67)	62.22 (51.34–72.05)
−	34	56

Positive predictive value (PPV); Negative predictive value (NPV)

Figure 3.

Comparative sensitivity and specificity of ESAT-6 and CFP-10 versus sputum AFB smear and Xpert MTB/RIF

Sensitivity in BC-PTB versus CD-PTB

ESAT-6 sensitivities were 84.8% in BC-PTB and 71.1% in CD-PTB (P > 0.05). CFP-10 sensitivities were 69.5% and 64.4%, respectively (P > 0.05), indicating comparable performance irrespective of bacteriological status (Table 3).

Table 3.

Sensitivity of ESAT-6 and CFP-10 in bacteriologically confirmed (BC-PTB) vs. clinically diagnosed (CD-PTB) pulmonary tuberculosis groups

Antigen	Sensitivity of BC-PTB group (number of positives/total cases)	Sensitivity of CD-PTB group (number of positives/total cases)	p value
ESAT-6	84.75% (50/59)	71.11% (32/45)	> 0.05
CFP-10	69.49% (41/59)	64.44% (29/45)	> 0.05

Combined antigen analysis

Parallel testing (positive if either antigen exceeded its cutoff) yielded a sensitivity of 94.2% (95% CI 87.4–97.6) with a negative predictive value of 90.5%, missing only six PTB cases. Tandem testing (positive only when both antigens were positive) increased specificity to 95.8% (95% CI 88.1–98.6) with a positive predictive value of 94.8% (Table 4).

Table 4.

Diagnostic performance of parallel and tandem testing strategies using BALF ESAT-6 and CFP-10 antigens

Detection strategy	Antigen combination	Case grouping	Performance indicators	Results (95% CI)
Parallel connection	ESAT-6 + CFP-10	Tuberculosis group	Sensitivity/missing rate	94.23% (87.36–97.63%)/5.77%
In series connection (electricity)	ESAT-6 + CFP-10	Non-tuberculosis group	Specificity/misdiagnosis rate	95.83% (87.50–98.92%)/4.17%

Discussion

ESAT-6 and CFP-10 are pivotal virulence factors secreted via the ESX-1 system. Their heterodimer enables immune evasion by binding toll-like receptors and CD91 on host phagocytes, activating MAPK/ERK signaling, suppressing lysosomal maturation, and promoting intracellular survival of Mycobacterium tuberculosis (MTB) [15]. Because the ESX-1 locus is absent from most non-tuberculous mycobacteria (NTM) and from Bacillus Calmette–Guérin (BCG), ESAT-6 and CFP-10 provide high species specificity [16,17–19]. Current immunodiagnostic assays—including interferon-γ release assays and ESAT-6/CFP-10 skin tests—leverage this specificity to avoid BCG or NTM cross-reactivity.

Consistent with prior studies [20–22], direct quantification of ESAT-6 and CFP-10 in BALF significantly improved diagnostic sensitivity compared to AFB smear and Xpert MTB/RIF. For instance, one investigation reported elevated sensitivity and specificity for these antigens in sputum culture supernatant (ESAT-6: 95.4%, 100%; CFP-10: 81.6%, 92.2%), attributable to in vitro bacterial amplification and selective MTBC growth [20]. Another study on tuberculous lymphadenitis demonstrated IHC sensitivity/specificity of 87.5%/85.3% for ESAT-6 and 75.0%/89.7% for CFP-10 [22]. Although our specificity was comparatively lower, parallel or tandem antigen testing effectively balanced sensitivity and specificity to serve distinct clinical purposes—either ruling out or confirming TB. While serum-based antigen detection has been attempted [21], low analyte levels due to systemic dilution limit its utility. Sputum, though direct, is often compromised by viscosity and inhibitors. In contrast, BALF sampling targets the alveolar infection site directly, yielding higher antigen concentrations and superior sensitivity, particularly in paucibacillary or sputum-scarce patients [23].

Antigen sensitivities were comparable in BC-PTB and CD-PTB groups, suggesting that secreted antigen detection is less dependent on bacillary burden than microbiological methods. Secreted proteins can diffuse beyond bacilli and may persist after bacterial lysis, thus complementing culture-based tests, particularly in paucibacillary disease [24–31].

Limitations include the absence of healthy controls—bronchoscopy precluded their inclusion—and the single-center design with a modest sample size. Multicenter studies are warranted to validate cutoff values and standardize protocols. The specificity reported herein may overestimate real-world performance, as our control group (IGRA-negative patients with tumors/non-TB infections) does not fully represent the spectrum of clinical conditions encountered during bronchoscopy. Future studies involving more heterogeneous control groups are needed to establish generalizable specificity estimates.

Conclusions

ELISA-based quantification of ESAT-6 and CFP-10 in BALF markedly enhances pulmonary tuberculosis (PTB) diagnostic sensitivity irrespective of bacteriological status. Parallel testing provides a reliable rule-out strategy, whereas tandem testing offers high specificity for confirming disease. BALF antigen detection is a promising adjunct to existing diagnostic algorithms, especially in smear- or Xpert-negative PTB.

Author contributions

Shiyu Fang and Jinxiong Zhuan were responsible for the ELISA assay of alveolar lavage fluid and wrote the main manuscript, Jie Sun was responsible for the collection of alveolar lavage fluid and the analysis of clinical data, and Fengjun Liu was responsible for the revision of the manuscript and experimental guidance. All authors reviewed the manuscript.

Funding

This study was supported by a grant from the Special Research Collaboration Project of North Sichuan Medical College (Grant No. CBY25-ZXZDA01).

Availability of data and materials

Data is provided within the manuscript or supplementary information files.

Declarations

Ethics approval and consent to participate

The study received ethical approval from the Ethics Committee of the Affiliated Hospital of Chuanbei Medical College (Approval No. 2023ER224-1). The rights and interests of the subjects were protected throughout the study. Participants provided written informed consent, and their privacy and confidentiality were strictly maintained. Data collected were used exclusively for this research and were not shared or used without authorization.

Consent for publication

All the participants provided written informed consent.

Competing interests

The authors declare no competing interests.