Detection of Mycobacterium tuberculosis complex in saliva by quantitative PCR: A potential alternative specimen for pulmonary tuberculosis diagnosis

Abstract

Background:

Current tuberculosis (TB) diagnostic tests primarily rely on sputum samples, yet many TB patients cannot produce sputum. This study explored whether saliva could be used instead of sputum to diagnose pulmonary TB (PTB).

Method:

The study included 32 patients with confirmed PTB and 30 patients with other respiratory diseases (ORD). Saliva from all study participants was subjected to quantitative (qPCR) assays targeting the IS1081 gene for detection of M. tuberculosis complex DNA.

Results:

The sensitivity of saliva IS1081 qPCR was 65.6 % (95 % CI 48.4–80.2 %) with positive results for 21/32 PTB cases, while the specificity was 96.7 % (95 % CI 85.9–99.6 %) with negative results for 29/30 participants with ORD. Sensitivity improved to 72.4 % (95 % CI 54.6–86.0 %) when sputum-Xpert was used as the reference standard, while remaining similar at 65.5 % (95 % CI 47.4–80.7 %) when culture was used as the reference standard. In receiver operating characteristic (ROC) curve analysis, the area under the curve (AUC) for saliva IS1081 qPCR was 82.5 % (95 % CI 71.7–93.3 %).

Conclusion:

Saliva testing offers a promising alternative to sputum for TB diagnosis among confirmed PTB cases. Larger multicenter studies, encompassing diverse clinical TB characteristics, are needed to provide improved estimates of diagnostic sensitivity and specificity.

1. Introduction

Tuberculosis (TB) is a leading cause of morbidity and mortality from a single infectious agent, despite being preventable and curable. An estimated 10.6 million new cases and 1.3 million deaths were attributed to TB in 2022 [1]. More than 3 million TB cases worldwide were estimated to remain undetected and unreported in 2022 [1], mainly due to inadequate diagnostic tools, which underscores the need to strengthen the capacity to diagnose the disease. Sputum is the specimen of choice for both conventional (microscopy and culture) and molecular-based TB diagnosis. However, it can be difficult or even impossible to obtain sputum from all TB patients, and such scarcity of sputum can contribute to delays in diagnosing TB patients, potentially resulting in increased mortality [2]. In addition, sputum-based TB diagnosis has reduced sensitivity in children because they are frequently unable to expectorate sputum and because of the paucibacillary nature of childhood TB [3]. Even in the group of TB patients who can provide sputum, commonly used sputum smear microscopy has limited sensitivity and misses the diagnosis in over one-third of patients seeking care [4]. Moreover, the gold standard to diagnose TB by culture has a prolonged incubation period (2–6 weeks) [5] that further delays TB diagnosis. Taken together, delays in TB diagnosis associated with the current sputum-based TB diagnosis approach have an impact on both disease prognosis at the individual level and transmission within the community [6].

Saliva is a very appealing specimen option for pulmonary TB (PTB) diagnosis as it has several advantages over sputum samples, including being easier and less invasive to collect, less infectious, less viscous, and more convenient for mass screening [7]. Saliva and oral swabs have been investigated in several research studies as a potential alternative for TB diagnosis; however, results of individual molecular-based studies have been variable, with a sensitivity range of 38–90 % and a specificity range of 92–100 % [8–14] using different modalities of molecular assays, including in-house quantitative PCR (qPCR), GeneXpert MTB/RIF (Xpert), and GeneXpert MTB/RIF Ultra (Xpert Ultra).

The insertion sequence (IS) 6110 is specific for the Mycobacterium tuberculosis complex (MTBC) [15] and a common target for in-house qPCR-saliva-based TB diagnosis studies [8,9]. Although IS6110 is mostly present in multiple copies in the M. tuberculosis genome [16], and therefore considered an ideal target for qPCR-based TB diagnosis, some M. tuberculosis strains have been reported to be devoid of IS6110 [17], which can result in false-negative results when IS6110 is the only target in such qPCR assays. IS1081 is an alternative target sequence for this kind of assay and, unlike the IS6110, it is present in six stable copies between M. tuberculosis lineages as well as across all M. tuberculosis complex (MTBC) species analyzed [18]. Currently, available molecular TB diagnostic tests, including Xpert and Xpert Ultra, are primarily optimized for testing sputum rather than saliva samples, and prior studies that tested saliva and oral swabs using Xpert and Xpert Ultra reported suboptimal sensitivity [10,11,13]. Therefore, we undertook the current study using an in-house IS1081 targeted qPCR assay to determine whether saliva can be used instead of sputum to diagnose PTB.

2. Method

2.1. Study population and setting

Adults aged between 18 and 70 years who presented with respiratory symptoms compatible with TB, including cough plus at least one of the following symptoms - fever, weight loss, hemoptysis, and night sweats were included in this study. Participants were prospectively recruited between January 2017 and August 2018 from two sites; Wereda 23 and Teklehaymanot Health Centers in Addis Ababa, Ethiopia. The participants were categorized into two groups: Group 1 comprised individuals with bacteriologically confirmed PTB who tested positive by sputum-Xpert and/or sputum-culture [19] while Group 2, the control group, consisted of individuals who presented with respiratory symptoms but tested negative for sputum-Xpert and sputum-culture. These patients responded to broad-spectrum antibiotics and other treatments and were ultimately diagnosed and treated for other respiratory diseases (ORD) by their physicians. Henceforth, this group will be referred to as ORD. A flow chart of the patient recruitment, sample analysis (diagnostic tests), and molecular detection (by IS1081 qPCR) is provided in Fig. 1.

Fig. 1.

Flow diagram of the study design used to evaluate saliva IS1081 qPCR to identify patients with bacteriologically confirmed pulmonary TB (PTB), compared to control participants with other respiratory diseases (ORD). Figure created with BioRender.com.

2.2. Demographic data, clinical data, and sample collection

The demographic and clinical data of the study participants were collected using structured questionnaires. Two sputum samples, one on the spot and one in the morning, were collected by asking the participant to cough and collecting the expectorated sputum in a sterile 50 mL Falcon tube. Subsequently, saliva was collected using a Salivette collection tube (Sarstedt, Nümbrecht, Germany). The Salivette tubes contained a cotton swab that was placed in each participant’s mouth with sterile pick-up forceps. Participants were asked to chew the swab for 1 min to stimulate salivation. The swab was then collected using pick-up forceps and placed in a Salivette tube. HIV testing was performed after pretest counseling as a part of routine care at the TB clinic.

2.3. Sample processing

The samples were transported on ice (up to +4 °C) from the recruitment site to the Armauer Hansen Research Institute (AHRI) laboratory within 2 h after collection. TB diagnosis was initially made by Xpert at a diagnostic clinic and then cultured at the AHRI laboratory to further confirm the diagnosis. Xpert testing (Cepheid, Sunnyvale, CA) of sputum was performed by mixing sputum with sample buffer at a ratio of 1:2 and processed and tested according to the manufacturer’s instructions. In the Biosafety Level 3 (BSL3) laboratory, Salivette tubes were centrifuged for 2 min at 1000×g to collect the saliva at the bottom of the tube; the saliva was then transferred into a cryotube and kept at −80 °C until DNA extraction was performed. Mycobacterial culturing was performed on sputum samples following the procedure indicated in the Mycobacteriology Laboratory Manual [20]. In brief, samples were decontaminated by the standard N-acetyl-L-cysteine and sodium hydroxide (NALC/NaOH) method with a final NaOH concentration of 1 %. An equal volume of standard NALC/NaOH solution was added to the specimen and incubated for 15 min. After neutralization with PBS and 15 min centrifugation at 3000×g, the sediment was re-suspended in 1 mL of sterile PBS and inoculated on two Löwenstein-Jensen (LJ) medium slopes and incubated at 37 °C. The LJ slopes were examined weekly for up to eight weeks for any visible growth. Positive culture was confirmed by Ziehl-Neelsen staining and smear microscopy. In parallel, the decontaminated sample volumes that remained after LJ inoculation were utilized for Ziehl-Neelsen staining and smear microscopy. Positive smears were semi-quantitatively graded as follows: “scanty” if the sputum contained 1–9 acid-fast bacilli (AFB) in 100 fields, “1+” for 10–99 AFB in 100 fields, “ 2+” if there were 1–10 AFB per field, and “3+” for more than 10 AFB per field [21].

2.4. DNA extraction from saliva

DNA was extracted from stored saliva samples. Each cryotube containing saliva was defrosted and then heated to 95 °C for 10 min to inactivate any pathogens. Five hundred microlitres of saliva was used for genomic DNA (gDNA) extraction using the QIAamp^® DNA Mini kit (Qiagen) following the manufacturer’s protocol, eluted in a volume of 50 μL, and stored at −20 °C until qPCR analysis.

2.5. IS1081 qPCR analysis

The primers and probe were designed to target the IS1081 insertion sequence, which is conserved among MTBC species [18,22]. The specificity of the primers and probe for the MTBC were tested in silico with BLAST software (www.ncbi.nlm.nih.gov/tools/primer-blast) and later confirmed by qPCR amplification of purified DNA from a range of MTBC species. The MTBC species used for primer/probe validation included M. tuberculosis Lineages I-4,7, M. africanum L5 and L6, M. canettii, M. microti, M. pinnipedii, M. caprae, M. bovis, and M. orygis. Their corresponding IS1081 qPCR results are shown in Supplementary Table 1. Both primers and probe were obtained from Eurofins Genomic India Pvt. Ltd. The reaction mixture for the IS1081 qPCR assay was 2.5 μL of 10 μM IS1081 forward primer 5′-GATCCTTCGAAACGACCA-3′, 2.5 μL of 10 μM IS1081 reverse primer 5′- CGGTGTCGATAAGATGAGA-3′, 0.5 μL 10 μM IS1081 probe [6FAM]-CGAAGGAAATGACGCAATGACCTC-[BHQ1], 12.5 μL of TaqMan^® Environmental Master Mix 2.0 (Applied Biosystems, [Thermofisher]), 2 μL nuclease-free water and 5 μL template DNA in a final volume of 25 μL. The qPCR cycling conditions were as follows: 50 °C for 2 min, 95 °C for 10 min, then 40 cycles of denaturation at 95 °C for 15 s, followed by annealing and extension at 58 °C for 1 min. Reactions were performed using the Biorad CFX96 Touch Real-Time PCR Detection System. All samples were tested in duplicate. A sample was considered positive if the average Ct value was less than 39 (with amplification in both reactions). For each qPCR assay round, purified gDNA of M. tuberculosis H37Rv and sterile molecular-grade water were used as positive and negative controls, respectively.

2.6. IS1081 qPCR assay: limit of detection

To assess the linearity, efficiency, and limit of detection (LoD) of the IS1081 qPCR assay, astandard curve was developed using M. tuberculosis H37Rv genomic DNA (gDNA). The initial gDNA stock concentration was 10 ng/μL and the subsequent ten-fold serial dilutions were performed with molecular grade water. The diluted samples were then subjected to qPCR analysis targeting IS1081. The LoD was defined as the lowest concentration at which amplification of both duplicate reactions was observed, below the predefined cut-off value (Ct < 39).

2.7. Data analysis

Frequencies and cross-tabulations were used to summarize descriptive statistics. Categorical variables were compared using the chi-square and Fisher’s exact tests, as appropriate. Sensitivity was calculated as the percentage of saliva IS1081 qPCR positive results among patients with bacteriologically confirmed PTB, whereas specificity was determined as the percentage of patients tested negative by saliva IS1081 qPCR among those diagnosed with ORD. Receiver operating characteristic (ROC) curve analysis was used to determine the area under the curve (AUC). The Ct value cut-off for defining positivity by IS1081 qPCR was established by selecting the value with maximal Youden’s index from ROC analysis. We estimated precision using exact binomial 95 % confidence intervals (CI). A p-value <0.05 was considered statistically significant. Statistical analyses were conducted using SPSS software (version 27.0; SPSS Inc., Chicago, IL, USA) or GraphPad Prism (version 8, GraphPad Software, San Diego, CA, USA).

2.8. Ethical consideration

Written informed consent was obtained from all the study participants. The study was approved by the AHRI/ALERT Ethics Review Committee (AAERC) (P021/16).

3. Result

3.1. Demographic and clinical characteristics of the study population

This study included 32 patients with bacteriologically confirmed PTB and 30 patients with ORD. The two groups showed no statistically significant differences in age or gender distribution (p > 0.05). A higher percentage of patients with active PTB had low BMI (53.1 %) compared to those with ORD (23.3 %), but the difference was not statistically significant (p = 0.063). Weight loss (p < 0.001) and smoking (p = 0.012) were significantly more prevalent in the active PTB group compared to the ORD group. No differences between the groups were found in the number of individuals with BCG vaccination, previous history of TB, or being HIV-positive (p > 0.05). Detailed demographic and clinical characteristics of the study population are given in Table 1.

Table 1.

Demographic and clinical characteristics of study participants.

		All Study participants (N = 62)	Patients with PTB (N = 32)	Controls (N = 30)	P Value
Age	Median (IQR)	37 (28.8–48.5)	36 (29.0–43.8)	40 (26.0–54.0)	0.125
Sex	Male	35 (56.5 %)	20 (62.5 %)	15 (50.0 %)	0.443
	Female	27 (43.5 %)	12 (37.5 %)	15 (50.0 %)
BMI	<18.5	24 (38.7 %)	17 (53.1 %)	7 (23.3 %)	0.063
	18.5–25	31 (50.0 %)	13 (40.6 %)	18 (60.0 %)
	>25–30	5 (8.1 %)	2 (6.3 %)	3 (10.0 %)
	>30	2 (3.2 %)	0 (0.0 %)	2 (3.2 %)
BCG scar	Present	26 (41.9)	15 (46.9 %)	11 (36.7 %)	0.719
Previous TB history	Yes	9(14.5 %)	4 (12.5 %)	5 (16.7 %)	0.861
Ever smoked	Yes	10 (16.4 %)	9 (29.0 %)	1 (3.3 %)	0.012
HIV	Positive	8 (12.9 %)	6 (18.8 %)	2 (6.7 %)	0.388
TB indicative symptoms	Cough >2 weeks	55 (88.7 %)	28 (87.5 %)	27 (90.0 %)	1.000
	Fever	19 (35.8)	11 (39.3 %)	8 (32.0)	0.775
	Night sweating	31 (50 %)	17 (53.1 %)	14 (46.7 %)	0.220
	Chest pain	34 (54.8 %)	19 (59.4 %)	15 (50 %)	0.610
	Loss of appetite	34 (54.8 %)	21 (65.6 %)	13 (43.3 %)	0.125
	Weight loss	25 (40.3)	21 (65.6 %)	4 (13.3 %)	<0.001
Sputum culture	Positive	29 (46.8 %)	29 (90.6 %)	0 (0.0 %)	<0.001
Sputum Xpert	Positive	29 (46.8 %)	29 (90.6 %)	0 (0.0 %)	<0.001
Sputum smear	Positive	18 (29.0 %)	18 (56.3 %)	0 (0.0 %)	<0.001
Smear grade*	Scanty	2 (3.2 %)	2 (6.3 %)
	1+	3 (4.8 %)	3 (9.4 %)
	2+	5 (8.0 %)	5 (15.6 %)
	3+	8 (12.9 %)	8 (25.0 %)

PTB: Pulmonary tuberculosis; IQR: Inter-quartile range; BCG: Bacille Calmette-Guérin; HIV: Human Immunodeficiency Virus; Scanty: 1–9 acid-fast bacilli (AFB) in 100 fields; 1+: 10–99 AFB in 100 fields; 2+: 1–10 AFB per field; 3+: >10 AFB per field.

Among the 30 patients with ORD, 11 (36.7 %) had bronchitis, 9 (30.0 %) had pneumonia, 3 (10.0 %) had upper respiratory tract infection (URTI), 2 (6.7 %) had allergic rhinitis, 1 (3.3 %) had allergic rhinitis with bronchitis, 1 (3.3 %) had URTI with pneumonia, 1 (3.3 %) had hilar lymphadenopathy, and the diagnosis for 2 (6.7 %) was not specifically recorded.

All active PTB cases were bacteriologically confirmed using sputum as the source of the sample. Among 32 PTB cases, 26 (81.3 %) were positive for both culture and Xpert tests whereas three (9.4 %) culture-positive participants were Xpert-negative, two (6.3 %) and one (3.1 %) Xpert-positive participants had negative and contaminated cultures, respectively. Additionally, 18 out of 32 PTB cases (56.3 %) were AFB smear microscopy positive, including eight (25.0 %) with an AFB microscopy smear grade of 3+, five (15.6 %) with a grade of 2+, three (9.4 %) with a grade of 1+, and two (6.3 %) with scanty grade.

3.2. Diagnostic performance of the saliva IS1081 qPCR assay

The overall IS1081 qPCR sensitivity for the detection of M. tuberculosis in saliva was 65.6 % (95 % CI 48.4–80.2 %), with positive results for 21 out of 32 bacteriologically confirmed PTB cases. The specificity was 96.7 % (95 % CI 85.5–99.6 %) with negative results for 29 out of 30 participants with ORD. Sensitivity improved to 72.4 % (21/29, 95 % CI 54.6–86.0 %) when sputum-Xpert was used as the reference independently (Table 2), while remaining similar at 65.5 % (19/29, 95 % CI 47.4–80.7 %) when culture was used as the reference standard independently. In the ROC curve analysis, the AUC for saliva IS1081 qPCR for the detection of bacteriologically confirmed PTB was 82.5 % (95 % CI 71.7–93.3 %) (Fig. 2).

Table 2.

Diagnostic accuracy of saliva IS1081 qPCR assay.

	Sputum Xpert reference standard (n/N, 95 % CI)	Sputum Xpert and/or sputum culture reference standard (n/N, 95 % CI)
Sensitivity a	21/29, 72.4 % (54.6–86.0)	21/32, 65.6 % (48.4–80.2)
Specificity b	29/30, 96.7 % (85.9–99.6)	29/30, 96.7 % (85.9–99.6)
PPV	21/22, 95.5 % (80.7–99.5)	21/22, 95.5 % (80.7–99.5)
NPV	29/37, 78.4 % (63.3–89.2)	29/40, 72.5 % (57.5–84.4)

^aNumber of patients with a positive saliva IS1081 assay/number of patients diagnosed with PTB.

^bNumber of control participants (patients with other respiratory diseases) with negative saliva IS1081 assay/number of control participants; CI: confidence interval.

Fig. 2.

Receiver operating characteristic (ROC) curve analysis showing Area Under the Curve (AUC) analysis of saliva IS1081 qPCR to identify patients with bacteriologically confirmed pulmonary TB, compared to control participants with other respiratory diseases. CI: confidence interval.

The sensitivity of saliva IS1081 qPCR was higher among smear microscopy-positive participants (83.3 %; 95 % CI 61.9–95.1 %) than smear microscopy-negative patients (42.9 %; 95 % CI 20.3–68.1 %) (p = 0.027). Of the six HIV-coinfected TB patients included in this study, four had positive results in the qPCR assay. No significant associations were observed between salivary IS1081 positivity and other factors, including sex, age, and BMI (data not shown).

3.3. Limit of detection of IS1081 qPCR assay

Ten-fold serial dilutions of H37Rv gDNA were used to assess the linearity, efficiency, and LoD of the IS1081 qPCR assay. The qPCR efficiency was 98.4 %, with R² = 0.998 (Supplementary Fig. 1). The assay remained reproducibly positive at template DNA dilutions between 1.0 and 1.0 × 10⁻⁵ ng/μL M. tuberculosis H37Rv gDNA, with only one of two duplicate reactions positive at 1.0 × 10⁻⁶ ng/μL, indicating a LoD between 1.0 × 10⁻⁵ and 1.0 × 10⁻⁶ ng/μL. Based on the mass of one genome of H37Rv (4,411,532 bp = 4.5170 fg) [23], this equates to between 1.1 and 11.1 genome equivalents (GE) per reaction.

3.4. Saliva IS1081 qPCR assay negative samples from PTB group

Eleven out of 32 (34.4 %) confirmed TB cases had IS1081 qPCR negative saliva. Three of these eleven TB patients were positive for sputum-smear microscopy, -culture, and -Xpert; four were sputum-smear microscopy-negative but positive for sputum-Xpert and sputum-culture; three were negative for sputum-Xpert and -smear microscopy but positive for sputum-culture; and one was sputum-Xpert positive but sputum-microscopy and -culture-negative.

4. Discussion

Effective TB transmission control strategies depend on the ability to diagnose the disease accurately and rapidly. Given the challenges associated with TB diagnosis in sputum-scarce PTB patients, there is increasing interest in testing alternative specimens that are easily collected from all TB patients. In the present study, we applied IS1081 qPCR on saliva samples and achieved an overall sensitivity of 65.6 % (95 % CI 48.4–80.2 %), This sensitivity is higher than some previously reported sensitivities for saliva/oral swab-based Xpert and Xpert Ultra assays, which ranged from 22.0 % to 56.4 % [10,11,13,24,25]. However, other studies have reported saliva/oral swab-based Xpert Ultra sensitivities ranging from 77.8 % to 90 %, which is higher than the results reported here [12,14,24]. Differences in the clinical characteristics of the study participants, or sampling and sample processing methods used, could explain the observed diagnostic performance of saliva-based TB diagnosis.

Previous studies using an IS6110 qPCR assay for M. tuberculosis detection from oral swabs have reported variable sensitivities and specificities ranging from 90 to 93 % and 92–100 %, respectively [8,9]. The IS6110 insertion sequence is present in variable copy numbers among different lineages of M. tuberculosis [26]. As a result, the sensitivity of the IS6110 targeted qPCR assay partly depends on the copy-number variability of the M. tuberculosis strains circulating in a specific geographical area. In contrast, IS1081 is uniformly present in six copies among M. tuberculosis lineages as well as across MTBC species [18,27]. Therefore, the results reported here are more likely to reflect the true sensitivity of saliva-based TB diagnosis among bacteriologically confirmed PTB patients, regardless of strain variability within a geographical area. Here, we reported a specificity of 96.7 % (95 % CI 85.9–99.6 %), that is, one false-positive result among 30 subjects with ORD. This particular subject was diagnosed with bronchitis during enrollment. We suspected that this patient may have had incipient TB at the time of enrollment. However, these could also be false positive results, for example, due to cross-contamination.

As saliva can host several pathogens apart from TB, it has been actively investigated for different infectious disease-causing pathogens, including Helicobacter pylori [28], Herpes Simplex Virus [29], and SARS-CoV-2 [30]. It has been reported that using saliva to diagnose SARS-CoV-2 has the advantage of being more practical for mass screening, reducing aerosol exposure for healthcare workers, and reducing the need for personal protective equipment because it is self-collected [31]. This could make it possible to diagnose multiple infectious diseases from a single saliva sample, especially respiratory diseases with overlapping clinical signs and symptoms. The SARS-CoV-2 pandemic was a huge setback in the fight against TB. In fact, the pandemic has been estimated to have resulted in a 5-year setback in terms of mortality from TB and a 9-year setback in terms of TB detection [32]. During the pandemic, it would have been beneficial to screen patients for TB and COVID-19 from a single saliva sample, as at least some missing TB cases would have been identified. In this study, we could not explore the possibility of detecting dual infection from saliva as our samples were collected before the pandemic. However, in the future, it may be interesting to investigate the possibility of using a single saliva sample to diagnose multiple infectious diseases.

The TB target product profile prioritized by the WHO suggests that acceptable non-sputum diagnostic tests should have a minimum overall sensitivity of >65 %. However, the sensitivity should exceed 98 % among patients with smear microscopy-positive PTB, while also maintaining a specificity of 98 % [33]. Based on these criteria, the diagnostic performance of saliva-based assays, reported by previous studies [12,14, 24] and by us here, can be considered as a good starting point to recognize the potential of saliva-based TB diagnosis approaches. Furthermore, compared to sputum, saliva may be more easily adapted to different point-of-care diagnostic approaches under development, as it is less viscous and more uniform in composition.

A key strength of this study is the enrolment of both control participants and TB cases from the same TB endemic areas. More importantly, the control participants in this study were patients with respiratory tract infections other than TB. This allowed us to evaluate the specificity of the saliva-based qPCR assay in a high TB endemic area. However, our study does have some limitations. First, all TB cases included in this study were bacteriologically confirmed. Therefore, diagnostic sensitivity might not be the same for clinically diagnosed TB patients with low bacterial load. Further evaluation is required to assess the test performance in paucibacillary TB cases. Second, culture and Xpert tests were conducted on fresh sputum samples, whereas the saliva IS1081 qPCR assay was conducted on frozen saliva samples. However, we anticipate minimal impact on test performance since the saliva samples were stored at −80 °C without undergoing any freeze-thaw cycles, thereby minimizing potential DNA degradation. Third, our qPCR assay could not detect drug resistance as it was solely designed for the detection of MTBC species. Finally, the small sample size hinders us from making more precise estimates of assay performance. Larger studies would provide improved estimates of diagnostic performance and may identify alternative assay cut-off values, depending on the intended diagnostic application.

5. Conclusion

The finding of this study showed that the DNA of MTBC can be detected in the saliva of the majority of bacteriologically confirmed PTB patients. Larger multicenter studies, including patients with broad clinical characteristics, are warranted to confirm the current findings. Efforts to develop an automated saliva-based molecular assay platform could further improve saliva-based TB diagnostic performance. The detection of M. tuberculosis DNA in the saliva of the majority of TB patients also demonstrates the potential for molecular drug resistance testing using saliva. If supported by larger studies, this could have a major impact on the global fight against TB, particularly if sensitivity is improved.