Oral washes and tongue swabs for Xpert MTB/RIF Ultra-based tuberculosis diagnosis in people with and without the ability to make sputum

Oral washes and tongue swabs for Xpert MTB/RIF Ultra-based tuberculosis diagnosis in people with and without the ability to make sputum | Research Square window.SnipcartSettings = { analytics: { enabled: false } }; (function() { var accessVector = localStorage.getItem('access_vector') || ''; window.dataLayer = window.dataLayer || []; if (accessVector) { window.dataLayer.push({ user: { profile: { profileInfo: { snid: accessVector } } } }); } })(); (function(w,d,s,l,i){w[l]=w[l]||[];w[l].push({'gtm.start':new Date().getTime(),event:'gtm.js'});var f=d.getElementsByTagName(s)[0],j=d.createElement(s),dl=l!='dataLayer'?'&l='+l:'';j.async=true;j.src='https://www.googletagmanager.com/gtm.js?id='+i+dl;f.parentNode.insertBefore(j,f);})(window,document,'script','dataLayer','GTM-K279D39R'); Browse Preprints In Review Journals COVID-19 Preprints AJE Video Bytes Research Tools Research Promotion AJE Professional Editing AJE Rubriq About Preprint Platform In Review Editorial Policies Our Team Advisory Board Help Center Sign In Submit a Preprint Cite Share Download PDF Research Article Oral washes and tongue swabs for Xpert MTB/RIF Ultra-based tuberculosis diagnosis in people with and without the ability to make sputum Loren Rockman, Shima Abdulgader, Stephanie Minnies, Zaida Palmer, and 12 more This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-6225530/v1 This work is licensed under a CC BY 4.0 License Status: Posted Version 1 posted You are reading this latest preprint version Abstract Background : Oral samples show promise for tuberculosis (TB) diagnosis. Data from different samples and people with sputum scarce TB are limited. Methods: We assessed Xpert MTB/RIF Ultra (Ultra) in symptomatic people at clinics (Cohort A, n=891) or at antiretroviral therapy (ART)-initiation without syndromic preselection (Cohort B, n=258). In Cohort A, we collected oral washes (OWs) and, separately, tongue swabs (flocked, foam with heat). In Cohort B, we collected OWs, three flocked tongue swabs (comparing one with heat to two pooled swabs) and, separately, buccal swabs, periodontal brushes. We offered sputum induction and did different culture methods on a subset of Cohort B tongue swabs. Results: In Cohort A, Ultra on OWs, flocked tongue and foam swabs had sensitivities of 80% (95% confidence interval 56, 94), 59% (53, 65) and 65% (58, 72) and high specificities. In Cohort B, OWs and single heated swabs had 71% (42, 92) and 64% (35, 87) sensitivity, respectively. Pooled tongue swabs, buccal swabs and periodontal brushes had low sensitivities. MGIT960 had the highest sensitivity [64% (35, 87)] of culture methods. Oral sampling detected TB in sputum-scarce people [Cohort A: 25% (7/28) flocked and foam swab-positive; Cohort B: 18% (10/56) OW-, 23% (13/56) single flocked swab-positive]. In Cohort B, this would at least double the people with a positive Ultra result (sputum or oral) if induction were unavailable. Conclusion: Ultra on OWs or foam tongue swabs has higher sensitivity than other oral-based approaches and detects sputum-scarce TB, resulted in more people diagnosed compared to Ultra on expectorated sputum. Molecular Biology Xpert MTB/RIF Ultra tuberculosis oral washes tongue swabs Figures Figure 1 Figure 2 Figure 3 Figure 4 Background Tuberculosis (TB), caused by Mycobacterium tuberculosis complex ( Mtb ), is a global pandemic (1). Molecular tests like Xpert MTB/RIF Ultra (Ultra; Cepheid, Sunnyvale, USA) and Truenat MTB Plus (Molbio Diagnostics, Goa, India) were designed for sputum. Non-sputum tests are urgently needed to reach the millions of people with TB undiagnosed each year (2-4). Tongue swabs, which collect biofilm from the dorsum tongue, have data to support use with the Ultra and Truenat MTB Ultima (Ultima) tests: Ultra has sensitivities from 72-75% with high specificity, with similar performance described for Ultima (5). Furthermore, discrete choice studies show tongue swabs are more acceptable than sputum to some people despite potential sensitivity trade-offs (6). Importantly, a test with diminished sensitivity can increase diagnostic yield if the number of individuals tested is higher (7, 8), however, there are few empiric data in the context of oral samples (9). Key gaps remain regarding the diagnostic potential of oral samples, including samples other than tongue swabs. Early proof-of-concept studies used in-house rather than commercially-available tests or, as tongue swabs are likely most beneficial in people unable to expectorate, paradoxically pre-selected people based on their ability to naturally produce sputum (10-13). Furthermore, the type of tongue swab (flocked, foam) and processing method may influence performance (14). Equally important, data from people with risk factors for TB not preselected based on symptoms, who are less able to make sputum and increasingly targeted as part of facility active case finding strategies, are scarce. More data are urgently needed to inform global policy. Data are also scarce regarding Mtb culturability from tongue swabs. One study found 44-58% of swabs to be MGIT960 culture-positive in people with TB (15). Culturable Mtb from the swab could reinforce microbiological reference standards for diagnostic accuracy evaluations and provide material for drug susceptibility testing. Lastly, culture methods that do not involve harsh NALC-NaOH decontamination and are designed for paucibacillary samples may have added utility. To address these knowledge gaps, we evaluated different types of oral samples, swabs, and processing methods for TB diagnosis in a HIV high burden setting where people were offered sputum induction. We hypothesized Ultra would detect Mtb in oral specimens with high concordance to sputum. Methods and Materials Ethics statement This work was approved by the Health Research Ethics Committee of Stellenbosch University (N14/10/136, M20/06/017, M20/06/018) and City of Cape Town (10570) . Written informed consent was obtained. Study Cohorts Cohort A: Adults (≥18 years) self-reporting with presumptive pulmonary TB symptoms meeting WHO criteria (at least one symptom for PLHIV, ≥2 for those without HIV) were recruited at clinics in Cape Town, South Africa. Cohort B: Antiretroviral therapy (ART) initiators (≥18 years), regardless of symptoms, were recruited in Cape Town as previously described (16, 17). Definitions Cohort A: People were classified as having TB using an extended microbiological reference standard (eMRS) if sputum was Ultra-positive or had Mtb -positive culture growth. Those categorized as not having TB had no positive sputum Ultra or culture results at least one negative result. Those missing a culture or Ultra result were, if other results negative, classified as not having TB. Cohort B: The eMRS consisted of two sputum cultures. If at least one culture was Mtb- positive, participants were classified as having TB. Participants with negative culture(s) did not have TB. Specimen collection All oral samples were collected prior to sputum collection after ≥30 min had passed since food or fluid was ingested and teeth brushed. Figure 1 summarises specimen collection and testing. Oral washes Each person (n=40 Cohort A, n=156 Cohort B) received two vials of 20 mL sterile water. One vial was used to rinse the mouth and discarded. The second vial’s contents were swirled in the mouth for a few seconds and spat into the vial. All samples were stored at -20 °C until processed. In Cohort A, OWs were collected from separate people than tongue swabs. In Cohort B, OWs were collected in everyone who gave tongue swabs and collection was done immediately after swab collection. Tongue swabs Cohort A: One dry flocked swab (FLOQSwabs code 520C; Copan Italia S.p.A., Brescia, Italy) followed by a dry foam swab (Medline Industries, Northfield, USA) were consecutively collected from 851 and 550 people, respectively, by scraping the tongue dorsum for 10-15 s and placing the swab in a dry tube (550 people had both swab types). As a control for each person, an air swab was collected in the same space as the participant by waving a foam swab (flocked prior to the start of foam swab sampling) in the air for 10-15 s immediately before sampling. Cohort B: One flocked swab was collected in 800 μL Tris-EDTA (TE) buffer from 156 people. The last 122 were asked to provide an additional two flocked swabs subsequently pooled into a single tube with 800 μl TE buffer. A flocked air swab was done as for Cohort A. In both cohorts, samples were stored (-20 °C) until processed. Buccal swabs and periodontal brushes In Cohort B, paired buccal swabs and periodontal brushes (n=102) were collected ( Supplementary text pg. 2 ) and stored (-20 °C) until processed. Collection occurred prior to the start tongue swab and OW collection. Sputum Cohort A participants were each asked to provide two sputa, while Cohort B participants were each asked to provide three sputa. Sputa were used for Ultra and MGIT960 culture [with 1% NALC-NaOH decontamination and MTBC confirmation using MTBDR plus (Bruker-Hain Diagnostics, Nehren, Germany); one culture in Cohort A, two in Cohort B]. Sputum induction was done (16, 18), however, it was only recorded whether people definitively required induction to make at least one ≥1 ml sputum in 584 Cohort A people and the 156 Cohort B people that gave tongue swabs. Figure 1 summarises specimens collected by Cohort. Specimen processing and testing All processing was performed in a biosafety level 3 (BSL3) laboratory. Oral wash testing using Ultra (Cohorts A and B) In Cohort A, OWs were concentrated, decontaminated, and processed with sample reagent (SR; Cepheid) per Supplementary text pg. 2 . Decontamination was not done in Cohort B. Tongue swab testing using Ultra (Cohorts A and B) Cohort A: Swabs were removed from storage and placed into a heating block (100 °C, 10 min), after which TE buffer was added to 2 mL and the whole volume tested with Ultra (19) with no SR. Cohort B: Single flocked swabs in TE buffer were removed from storage and immediately boiled (100°C, 10 min) followed by SR addition (1.6 mL to 800 μL sample) and Ultra (19). For the 122 people who also gave a double swab, no heating was done, 2:1 SR added, and Ultra done (Supplementary Figure 1). Air swab controls in people who had tongue swabs (Cohorts A and B) Air swabs was processed and tested with Ultra using the same procedure as tongue swabs every tenth patient. If a participant had a positive tongue swab their air swab was tested. Buccal swabs and periodontal brushes (Cohort B) Stored samples were removed, processed, and tested with Ultra using the same procedure as tongue swabs in Cohort B. Tongue swab culture (Cohort B) We did different types of culture (MGIT960, TiKa (20), early bactericidal activity (EBA) (21); methodology in Supplementary text pg. 2 ) on single flocked tongue swabs (without heat inactivation). Speciation on positive growth was done using Ultra on a concentrated MGIT960 tube (22). Analysis Methods and reporting are per STARD guidelines (23). Diagnostic accuracy metrics were calculated using Excel (Microsoft, Redmond, USA) and compared using prtest (24) in STATA version 16.0 (StataCorp, Texas, USA). The effect on sensitivity and specificity of Ultra trace results removed or reclassified to Ultra-negative in 2x2 tables was evaluated. Continuous data were compared with GraphPad Prism version 8.0.1 (GraphPad Software, San Diego, USA), also used for linear regression and correlation analysis. Diagnostic yield (DYT, diagnostic yield in those tested; DYD, diagnostic yield in those diagnosed) was calculated as described (8) and defined in Supplementary text pg. 3 . Morbidity score information (TBscoreII) was collected (25). Venn diagrams were made using Interactivenn (26). In Cohort B, people were designated asymptomatic based on the WHO four symptom screen (27). P-values ≤0.05 were significant. Unsuccessful results are those not positive or negative by any test. We compared SPC C T values from Ultra to measure inhibition (lower SPC C T s mean less inhibition) (28). Results Participant demographics People in Cohort A were, compared to Cohort B, more likely to be female, have higher morbidity, and more likely to have previous TB ( Table 1) . In both cohorts, people with TB were more likely to be male and have higher morbidity. Diagnostic accuracy of Ultra Data are summarised in Figures 2 and 3 . All air swabs were negative. Cohort A OWs: No unsuccessful results occurred. Sensitivity was 80% (56, 94) and specificity 80% (56-94). Sensitivity was higher among those without HIV compared to PLHIV [94% (70-100) vs 25% (1-81); p=0.002] ( Table 2 ). Four false-positive results occurred (all trace semi-quantitation, two previous TB). Tongue swabs: 3% (25/851) of flocked swabs and 2% (13/550; p=0.611) of foam swabs had unsuccessful results (3 both, 22 flocked only, 10 foam only; mostly overpressure errors). Flocked swabs had lower sensitivity than foam swabs [59% (53-65) vs. 65% (58-72); p=0.001] with high specificities [94% (91-96) vs. 92% (89-95); p=0.100] ( Table 2 ). Amongst false-positive swabs, 26% (9/34) of flocked swabs and 39% (11/28) of foam swabs were from people programmatically empirically treated (no positive bacteriology at treatment start). Different trace recategorization strategies resulted in small sensitivity decreases and large specificity increases ( Supplementary Table 1 ). Cohort B OWs: No unsuccessful results occurred. Sensitivity was 71% (42-92) and specificity 92% (86- 96). Of the 12 false positives, all were Ultra semi-quantitation category trace. 17% (2/12) were Ultra flocked tongue swab positive. Buccal swabs and periodontal brushes: No unsuccessful results occurred. Sensitivity was 7% (0-34) and 14% (2-43), respectively, with 98% (92-100) specificity for both ( Supplementary Table 2 ). Tongue swabs: 1% (2/156) of single swabs, and 1% (1/122; p=0.711) double swabs generated unsuccessful results. In head-to-head analyses, single vs. double swab sensitivity was 67% (30-93) and 22% (3-60; p=0.068) and specificity 82% (74-89) vs. 96% (91-99; p=0.009) ( Table 3 ). Alternative evidence of TB occurred in 23% (6/26) of false-positive single flocked swabs [two sputum Ultra-positive, one Alere Determine TB LAM Ag (Abbott, USA)-positive, three positive MGIT960 tongue swab culture] and 25% (1/4) of false-positive double swabs (sputum Ultra-positive). Different trace recategorization strategies resulted in small sensitivity decreases and large specificity increases ( Supplementary Table 3 ). Asymptomatic TB: 58% (150/258) of people were asymptomatic and, of these, 6% (9/150) had TB. 33% (3/9), 22% (2/9) and 0% (0/9) were positive using OW, single flocked swab and double flocked swabs, respectively. Diagnostic accuracy of tongue swab culture in Cohort B MGIT960 and TiKa had 64% (35-87) vs. 36% (13-65; p=0.131) sensitivity, while specificity was 88% (82-93) vs. 94% (89-98; p=0.060) (Supplementary Table 4 ). For comparison, sputum TiKa culture sensitivity and specificity were 79% (49-95; p=0.022 vs. TiKa on tongue swabs) and 97% (93-99; p=0.238), respectively. In the subset who underwent tongue swab EBA culture, sensitivity and specificity were 33% (4-78) and 95% (85-99). Yield of Ultra Figure 4 shows people who tested positive by Ultra on different samples, as well as sputum culture. Ultra yield metrics are compared in Supplementary Table 5 . Similar patterns for DYD occurred as those described for DYT. We also calculate yields from tongue swab culture (Supplementary text pg. 4) , which were low. Cohort A Overall: OW DYT (50%, 20/50) was like that for sputum. Foam tongue swabs had a higher DYT (27%, 150/550) point estimate than flocked tongue swabs (23%, 131/150) but this did not reach significance. Sputum DYT was significantly higher [33% (180/550] than flocked and foam swabs. If sputum induction were unavailable: No one with an OW had sputum induction information. In people with both tongue swab types, 5% (28/550) could not expectorate sputum. 25% (7/28) and 25% (7/28) were foam or flocked tongue swab positive. If tongue swabs were done in people who could not expectorate sputum, people bacteriologically diagnosed rapidly would increase from 175 (Ultra-positive on expectorated sputum) to 182 for flocked and foam swabs, [4% (2-6) increase]. Cohort B Overall: Amongst people with all three sample types, DYT point estimates were highest for single flocked tongue swab compared to OWs and double tongue swabs [21% (26/122), 14% (17/122), 5% (6/122)]. Sputum DYT was 12% (5/122). If sputum induction were unavailable: In people who had all three sample types (OWs, single and double swabs), 35% (43/122) could not expectorate sputum, of which 16% (7/43), 26% (11/43) and 5% (2/43) were positive on each oral sample type. If oral sampling was done in people who could not expectorate sputum, bacteriological diagnoses would change from 8 (Ultra-positive on expectorated sputum) to 15, 19, and 10 for OWs, single swabs, and double swabs, respectively [+88% (82-93), +138% (130-145) and 25% (17-33) respectively. If induction were unavailable, sputum DYT decreased to 7% (8/122), lower than OWs [14% 17/122; p=0.057] and single flocked swabs [21% (26/122); p=0.0009]. Ultra inhibition In Cohort A, oral washes had less PCR inhibition than sputum, however, this did not occur in Cohort B. No other differences of note occurred across sample types (Supplementary text pg. 4). Discussion Our key findings are: 1) OW Ultra had the highest sensitivity amongst the methods tested (71-80%), 2) Ultra on foam swabs had higher sensitivity than flocked swabs (65% vs. 59%; Cohort A), 3) Ultra on oral samples swabs diagnosed TB in many people who could not naturally expectorate, permitting yield to exceed that of sputum-based testing in Cohort B where sputum scarcity was more common, and 4) other approaches (Ultra on double tongue swabs, buccal swabs, and periodontal brushes) were suboptimal. These data suggest oral sampling, especially oral washes and foam swabs, can improve the diagnosis of TB, especially when sputum scarcity is accounted for. OW Ultra had a higher sensitivity point estimate than other approaches. To our knowledge, we report the first study using OW Ultra for TB diagnosis. Previous studies (29, 30) used in-house PCR methods applied to OW reported sensitivities of 77% and 88%, broadly our findings. We suggest OWs are included as a comparator in research on tongue swabs for TB diagnosis going forward as, unlike swabs, they require less processing. We observed a small sensitivity increase with foam swabs compared to flocked swabs, likely attributable to the larger amount of biomass bound to the foam swab (15, 31). Foam swabs have an added advantage in that they are relatively cheap (0.14 USD per swab) compared to flocked swabs. On a practical note, however, participants report foam swabs make their tongues feel “dry”. Furthermore, the lack of a breakpoint means swab heads need to be cut off, however, this is addressable. The magnitude of the sensitivity improvement from foam swabs may, vs. flocked swabs, be even greater in people with earlier stage disease. We showed Ultra on swabs can detect TB in people who cannot make. As TB testing programmatically expands, including to people with early stage potentially subclinical TB, the proportion of people with sputum scarce TB that require testing will increase. Thus, our study addresses a key gap: most studies on oral samples for TB diagnosis have either not recruited people with sputum scarce TB or have offered induction and not been able to disaggregate performance in people who, without induction, cannot expectorate. This is important because, even if tongue swabs perform well in sputum expectorators, it is hard to justify not testing sputum if it is available. Our data are thus novel in that they give performance data in a type of person most likely to benefit from non-sputum tests. For example, in Cohort B, the number of people with a positive Ultra result would at least double with the use of oral sampling in sputum scarce. We compared a single flocked swab with heat lysis, which had high sensitivity in earlier work using contrived samples (32), to a double flocked swab without heat lysis, which had lower sensitivity. This suggests heating is critical to release DNA. It remains to be seen if heating with a double swab improves sensitivity further, however, given our results from foam swabs, it is likely that steps that input material would be beneficial, providing heat lysis occurs. In addition to the double swab method, we evaluated other specimen types (buccal swabs, periodontal brushes), as well as different tongue swabs culture methods, however, neither were promising. Our study has strengths and limitations. Recruitment was programmatic in nature in that it included people with symptoms and non-symptom-based risk factors for which South African guidelines (33, 34) require molecular testing, even if they were not yet symptomatic and/or could not expectorate sputum. More data are needed in these groups where oral sample testing is likely to be most impactful (7). Different processing methods, such as those which maximise time to recover DNA buccal swabs and periodontal brushes were in buffer for short periods) should be further explored alongside OWs, which notably does not require heating. Lastly, the inclusion of people who cannot make sputum naturally and crucially the availability of induction information are strengths. In conclusion, Ultra on oral samples – especially foam tongue swabs and OW – is sensitive and highly specific and can significantly increase the overall number of people with a rapid positive bacteriological result when applied to people who cannot naturally make sputum. Declarations Funding: South African Medical Research Council, National Research Foundation Acknowledgements The authors thank the National Health Laboratory Services, Cape Town, South Africa, members of the Clinical Mycobacteriology & Epidemiology (CLIME) research group, and study participants. Research reported in this publication was supported by the South African Medical Research Council (SAMRC), the National Research Foundation (NRF) and the postgraduate bursary department of SU are hereby also acknowledged for funding this project. LR acknowledges funding from the SAMRC. The work reported herein was made possible through funding by the SAMRC through its Division of Research Capacity Development under the SAMRC Internship Scholarship Programme. The content hereof is the sole responsibility of the authors and does not necessarily represent the official views of the SAMRC. GT reports funding from the EDCTP2 programme supported by the EU (RIA2018D-2509, PreFIT; RIA2018D-2493, SeroSelectTB; RIA2020I-3305, CAGE-TB) and the National Institutes of Health (D43TW010350, U01AI152087, U54EB027049, and R01AI136894). Cepheid donated cartridges used in the study. They did not, however, have a role in study design or result interpretation. References World Health Organization. Global Tuberculosis Report 2023. World Health Organization, Geneva, Switzerland. 2023. World Health Organization. Target product profile for tuberculosis diagnosis and detection of drug resistance. World Health Organization, Geneva, Switzerland. 2024. Nathavitharana RR, Garcia-Basteiro AL, Ruhwald M, Cobelens F, Theron G. Reimagining the status quo: How close are we to rapid sputum-free tuberculosis diagnostics for all? EBioMedicine. 2022;78. Abdulgader SM, Okunola AO, Ndlangalavu G, Reeve BW, Allwood BW, Koegelenberg CF, et al. Diagnosing tuberculosis: what do new technologies allow us to (not) do? Respiration. 2022;101(9):797-813. Wood RC, Luabeya AK, Dragovich RB, Olson AM, Lochner KA, Weigel KM, et al. Diagnostic accuracy of tongue swab testing on two automated tuberculosis diagnostic platforms, Cepheid Xpert MTB/RIF Ultra and Molbio Truenat MTB Ultima. Journal of Clinical Microbiology. 2024;62(4):e00019-24. Shah K, Nalugwa T, Marcelo D, Nakawunde R, Trinh T, Shankar D, et al. A Multi-Country Evaluation of Patient Preferences for Tuberculosis Diagnostics: A Discrete Choice Experiment to Inform WHO Target Product Profiles. medRxiv. 2024:2024.06.19.24309124. de Nooy A, Ockhuisen T, Korobitsyn A, Khan SA, Ruhwald M, Ismail N, et al. Trade-offs between clinical performance and test accessibility in tuberculosis diagnosis: a multi-country modelling approach for target product profile development. The Lancet Global Health. 2024;12(7):e1139-e48. Broger T, Marx FM, Theron G, Marais BJ, Nicol MP, Kerkhoff AD, et al. Diagnostic yield as an important metric for the evaluation of novel tuberculosis tests: rationale and guidance for future research. The Lancet Global Health. 2024;12(7):e1184-e91. Bezuidenhout C, Long L, Nichols B, Meyer-Rath G, Fox MP, Olifant S, et al. USING SPUTUM AND TONGUE SWAB SPECIMENS FOR IN-HOME POINT-OF-CARE TARGETED UNIVERSAL TESTING FOR TB OF HOUSEHOLD CONTACTS: AN ACCEPTABILITY AND FEASIBILITY ANALYSIS. medRxiv. 2024. Valinetz ED, Cangelosi GA. A Look Inside: Oral Sampling for Detection of Non-Oral Infectious Diseases. Journal of Clinical Microbiology. 2021:JCM. 02360-20. Luabeya AK, Wood RC, Shenje J, Filander E, Ontong C, Mabwe S, et al. Noninvasive detection of tuberculosis by oral swab analysis. Journal of clinical microbiology. 2019;57(3):e01847-18. Nicol MP, Wood RC, Workman L, Prins M, Whitman C, Ghebrekristos Y, et al. Microbiological diagnosis of pulmonary tuberculosis in children by oral swab polymerase chain reaction. Scientific reports. 2019;9(1):10789. Wood RC, Luabeya AK, Weigel KM, Wilbur AK, Jones-Engel L, Hatherill M, et al. Detection of Mycobacterium tuberculosis DNA on the oral mucosa of tuberculosis patients. Scientific reports. 2015;5(1):1-5. Church EC, Steingart KR, Cangelosi GA, Ruhwald M, Kohli M, Shapiro AE. Oral swabs with a rapid molecular diagnostic test for pulmonary tuberculosis in adults and children: a systematic review. The Lancet Global Health. 2024;12(1):e45-e54. Wood RC, Andama A, Hermansky G, Burkot S, Asege L, Job M, et al. Characterization of oral swab samples for diagnosis of pulmonary tuberculosis. PloS one. 2021;16(5):e0251422. Reeve BW, Ndlangalavu G, Mishra H, Palmer Z, Tshivhula H, Rockman L, et al. Point-of-care C-reactive protein and Xpert MTB/RIF Ultra for tuberculosis screening and diagnosis in unselected antiretroviral therapy initiators: a prospective, cross-sectional, diagnostic accuracy study. The Lancet Global Health. 2024;12(5):e793-e803. Mann T, Gupta RK, Reeve BW, Ndlangalavu G, Chandran A, Krishna AP, et al. Blood RNA biomarkers for tuberculosis screening in people living with HIV before antiretroviral therapy initiation: a diagnostic accuracy study. The Lancet Global Health. 2024;12(5):e783-e92. Mishra H, Reeve BW, Palmer Z, Caldwell J, Dolby T, Naidoo CC, et al. Xpert MTB/RIF Ultra and Xpert MTB/RIF for diagnosis of tuberculosis in an HIV-endemic setting with a high burden of previous tuberculosis: a two-cohort diagnostic accuracy study. The Lancet Respiratory Medicine. 2020;8(4):368-82. Cepheid. Xpert MTB/RIF Ultra product brochure. 2017. Goosen WJ, Kleynhans L, Kerr TJ, van Helden PD, Buss P, Warren RM, et al. Improved detection of Mycobacterium tuberculosis and M. bovis in African wildlife samples using cationic peptide decontamination and mycobacterial culture supplementation. Journal of Veterinary Diagnostic Investigation. 2022;34(1):61-7. Diacon AH, Donald PR. The early bactericidal activity of antituberculosis drugs. Expert review of anti-infective therapy. 2014;12(2):223-37. Ghebrekristos YT, Beylis N, Centner CM, Venter R, Derendinger B, Tshivhula H, et al. Xpert MTB/RIF Ultra on contaminated liquid cultures for tuberculosis and rifampicin-resistance detection: a diagnostic accuracy evaluation. The Lancet Microbe. 2023;4(10):e822-e9. Cohen JF, Korevaar DA, Altman DG, Bruns DE, Gatsonis CA, Hooft L, et al. STARD 2015 guidelines for reporting diagnostic accuracy studies: explanation and elaboration. BMJ open. 2016;6(11). StataCorp L. prtest-Tests of proportions [cited 2023 Oct 22]. Available from: https://www.stata.com/manuals/rprtest.pdf. Rudolf F, Lemvik G, Abate E, Verkuilen J, Schön T, Gomes VF, et al. TBscore II: refining and validating a simple clinical score for treatment monitoring of patients with pulmonary tuberculosis. Scandinavian journal of infectious diseases. 2013;45(11):825-36. Heberle H, Meirelles GV, da Silva FR, Telles GP, Minghim R. InteractiVenn: a web-based tool for the analysis of sets through Venn diagrams. BMC bioinformatics. 2015;16(1):1-7. World Health Organization. WHO consolidated guidelines on tuberculosis. Module 2: screening-systematic screening for tuberculosis disease. 2021. Blakemore R, Nabeta P, Davidow AL, Vadwai V, Tahirli R, Munsamy V, et al. A multisite assessment of the quantitative capabilities of the Xpert MTB/RIF assay. American journal of respiratory and critical care medicine. 2011;184(9):1076-84. Davis JL, Huang L, Kovacs JA, Masur H, Murray P, Havlir DV, et al. Polymerase chain reaction of secA1 on sputum or oral wash samples for the diagnosis of pulmonary tuberculosis. Clinical infectious diseases. 2009;48(6):725-32. Evans D, Goyal M, Taylor I, Shaw R. Identification of M. tuberculosis ribosomal RNA in mouthwash samples from patients with tuberculosis. Respiratory medicine. 1994;88(9):687-91. Olson AM, Wood RC, Weigel KM, Yan AJ, Lochner KA, Dragovich RB, et al. High-sensitivity detection of Mycobacterium tuberculosis DNA in tongue swab samples. Journal of Clinical Microbiology. 2025;63(2):e01140-24. Andama A, Reza T, Jaganath D, Mulondo J, Semitala F, Wobudeya E, et al. Accuracy of Oral Swab Testing Using Xpert MTB-RIF Ultra for Tuberculosis Diagnosis. C106 FIRE TB LIVES ON: American Thoracic Society; 2022. p. A4928-A. Martinson NA, Nonyane BA, Genade LP, Berhanu RH, Naidoo P, Brey Z, et al. Evaluating systematic targeted universal testing for tuberculosis in primary care clinics of South Africa: A cluster-randomized trial (The TUTT Trial). Plos Medicine. 2023;20(5):e1004237. Department of Health SA. TB Screening and Testing Standard Operating Procedure,. 2022. Tables Table 1: Demographic and clinical characteristics. Data are n/N (%) or median (IQR). Cohort A Cohort B Overall (n=891) Definite TB (n=315) Non-TB (n=576) Overall (n=258) Definite TB (n=28) Non-TB (n=230) Demographic Age (years) 38 (30-48) 38 (30-47) 38 (30 - 48) p=0.364* 37 (31-44) p=0.956 € 36 (31-44) p=0.378 € 39 (31 -43) p=0.352* p=0.002 € Female 445/891 (50) 131/315 (42) 314/576 (55) p=0.002 * 156/258 (60) p=0.003 € 11/28 (39) p=0.813 € 145/230 (63) p=0.015 * p=0.027 € Clinical HIV-positive 355/891 (40) 136/315 (44) 219/576 (38) p=0.133* 258/258 (100) p<0.0001 € 28/28 (100) p<0.0001 € 230/230 (100) p<0.0001 € CD4 count (cells/μL) 305 (269-903) 582 (250-945) 305 (290-605) p=0.587* 293 (143-434) p<0.0001 € 59 (22-135) p<0.0001 € 299 (171 - 438) p=0.013 * p<0.0001 € TBScoreII 3 (2-3) 3 (3-4) 3 (2-3) p<0.0001 * 2 (0-2) p<0.0001 € 3 (2-4) p=0.001 € 1 (0-2) p=0.0006 * p<0.0001 € Previous TB 398/891 (45) 114/315 (36) 284/576 (49) p=0.0002 * 26/258 (10) p<0.0001 € 2/28 (7) p=0.002 € 24/230 (10) p=0.585* p<0.0001 € Within column p-values: *Definite TB vs. Non-TB, € Across cohorts for people of the same TB status Abbreviation: CD4, cluster of differentiation 4; HIV, human immunodeficiency virus; IQR, interquartile range Table 2: Diagnostic accuracy of Ultra on OW or tongue swabs compared to an eMRS for the detection of TB stratified by HIV status in Cohort A. Data are %, (95% CI), and n/N. Overall β (n=891) HIV-negative (n=526) HIV-positive (n=355) Sensitivity Specificity PPV NPV Sensitivity Specificity PPV NPV Sensitivity Specificity PPV NPV Oral wash 80 (56, 94) 16/20 80 (56, 94) 16/20 80 (56, 94) 16/20 80 (56, 94) 16/20 94 (70, 100) 15/16 93 (57, 96) 14/17 83 (59, 96) 15/18 93 (68, 100) 14/15 25 (1, 81) 1/4 p=0.002 * 67 (9, 99) 2/3 p=0.531* 50 (1, 99) 1/2 p=0.264* 40 (5, 85) 2/5 p=0.010* Flocked swab 59 (53, 65) 164/277 94 (91, 96) 515/549 83 (77, 88) 164/198 82 (79, 85) 515/628 56 (48, 64) 84/150 82 (90, 96) 307/329 79 (70, 87) 84/106 82 (78, 86) 307/373 63 (54, 71) 78/124 p=0.420* 95 (92, 98) 204/214 p=0.660* 89 (80, 94) 78/88 p=0.416* 82 (76, 86) 204/250 p=0.057* Foam swab 65 (58, 72) 123/188 p=0.039 € 92 (89, 95) 321/349 p=0.100 € 81 (74, 87) 123/151 p=0.679 € 83 (79, 87) 321/386 p=0.007 € 57 (47, 67) 58/101 p=0.385 € 82 (90, 97) 197/209 p=0.100 € 83 (72, 91) 58/70 p=0.617 € 82 (77, 87) 197/240 p=0.117 € 74 (63, 83) 62/84 p=0.164* p=0.013 € 89 (82, 94) 119/134 p=0.660* p=0.100 € 81 (70, 89) 62/77 p=0.314* p=0.906 € 84 (77, 90) 119/141 p=0.005* p=0.010 € Within row p-values: *HIV-negative vs. HIV-positive Within column p-values: € vs. flocked swab Abbreviations: CI, confidence interval; eMRS, extended microbiological reference standard; HIV, human immunodeficiency virus; NPV, negative predictive value; OW, oral wash; PPV, positive predictive value; Ultra, Xpert MTB/RIF Ultra β Ten people had unknown HIV status Table 3: Diagnostic accuracy of Ultra on OW or tongue swabs compared to a double sputum culture as an eMRS for TB detection in Cohort B. Data are %, 95% CI, and n/N. Head-to-head (n=122) Non-head-to-head (n=156) Sensitivity Specificity PPV NPV Sensitivity Specificity PPV NPV Oral wash 56 (21, 86) 5/9 89 (82, 94) 101/113 29 (10, 56) 5/17 96 (91, 99) 101/105 71 (42, 92) 10/14 p=0.435* 92 (86, 96) 130/142 p=0.556* 45 (24, 68) 10/22 p=0.307* 97 (93, 99) 130/134 p=0.725* Single boiled tongue swab 67 (30, 93) 6/9 p=0.065 Ω 82 (74, 89) 91/110 p=0.151 Ω 23 (9, 44) 6/26 p=0.695 Ω 97 (91, 99) 91/94 p=0.864 Ω 64 (35, 87) 9/14 p=0.907* p=0.705 Ω 81 (74, 87) 114/140 p=0.79* p=0.018 Ω 26 (12, 43) 9/35 p=0.813* p=0.143 Ω 96 (90, 99) 114/119 p=0.700* p=0.609 Ω Double tongue swab 22 (3, 60) 2/9 p=0.160 Ω p=0.068 ≠ 96 (91, 99) 108/112 p=0.042 Ω p=0.009 ≠ 33 (4, 78) 2/6 p=0.858 Ω p=0.639 ≠ 94 (88, 98) 108/115 p=0.480 Ω p=0.395 ≠ 22 (3, 60) 2/9 p=0.022 Ω p=0.048 ≠ 96 (91, 99) 108/112 p=0.117 Ω p=0.0004 ≠ 33 (4, 78) 2/6 p=0.595 Ω p=0.729 ≠ 94 (88, 98) 108/115 p=0.247 Ω p=0.526 ≠ Within row p-values: *Head-to-head vs. Non-head-to-head Within column p-values: Ω vs. OW, ≠ vs. single boiled swab Abbreviations: CI, confidence interval; eMRS, extended microbiological reference standard; NPV, negative predictive value; OW, oral wash; PPV, positive predictive value; Ultra, Xpert MTB/RIF Ultra Additional Declarations The authors declare no competing interests. Supplementary Files OralUltraSupplementaryresultsLR13March.docx Cite Share Download PDF Status: Posted Version 1 posted You are reading this latest preprint version Research Square lets you share your work early, gain feedback from the community, and start making changes to your manuscript prior to peer review in a journal. As a division of Research Square Company, we’re committed to making research communication faster, fairer, and more useful. We do this by developing innovative software and high quality services for the global research community. Our growing team is made up of researchers and industry professionals working together to solve the most critical problems facing scientific publishing. Also discoverable on Platform About Our Team In Review Editorial Policies Advisory Board Help Center Resources Author Services Accessibility API Access RSS feed Manage Cookie Preferences © Research Square 2026 | ISSN 2693-5015 (online) Privacy Policy Terms of Service Do Not Sell My Personal Information {"props":{"pageProps":{"initialData":{"identity":"rs-6225530","acceptedTermsAndConditions":true,"allowDirectSubmit":true,"archivedVersions":[],"articleType":"Research Article","associatedPublications":[],"authors":[{"id":428802408,"identity":"9ea4bacb-d632-4d90-88ab-b2ffa536a844","order_by":0,"name":"Loren Rockman","email":"","orcid":"","institution":"Stellenbosch University","correspondingAuthor":false,"prefix":"","firstName":"Loren","middleName":"","lastName":"Rockman","suffix":""},{"id":428802409,"identity":"9cff8a2b-5fc6-4b62-9a1f-25b47e7aec02","order_by":1,"name":"Shima Abdulgader","email":"","orcid":"","institution":"Stellenbosch University","correspondingAuthor":false,"prefix":"","firstName":"Shima","middleName":"","lastName":"Abdulgader","suffix":""},{"id":428802410,"identity":"931853b2-39b6-4a4f-a3b2-a7d75568abb2","order_by":2,"name":"Stephanie Minnies","email":"","orcid":"","institution":"Stellenbosch University","correspondingAuthor":false,"prefix":"","firstName":"Stephanie","middleName":"","lastName":"Minnies","suffix":""},{"id":428802411,"identity":"37a3d120-9495-431a-824c-1b7ec7c274e6","order_by":3,"name":"Zaida Palmer","email":"","orcid":"","institution":"Stellenbosch University","correspondingAuthor":false,"prefix":"","firstName":"Zaida","middleName":"","lastName":"Palmer","suffix":""},{"id":428802412,"identity":"6cdb46ed-dfa5-4c96-b734-a3b9228e53d2","order_by":4,"name":"Charissa C. Naidoo","email":"","orcid":"","institution":"Stellenbosch University","correspondingAuthor":false,"prefix":"","firstName":"Charissa","middleName":"C.","lastName":"Naidoo","suffix":""},{"id":428802413,"identity":"a716cf66-e94e-42e4-b5bf-2a993d23931b","order_by":5,"name":"Daphne Naidoo","email":"","orcid":"","institution":"Stellenbosch University","correspondingAuthor":false,"prefix":"","firstName":"Daphne","middleName":"","lastName":"Naidoo","suffix":""},{"id":428802414,"identity":"b02877b2-a7cb-47db-ad54-436a41620383","order_by":6,"name":"Rouxjeane Venter","email":"","orcid":"","institution":"Stellenbosch University","correspondingAuthor":false,"prefix":"","firstName":"Rouxjeane","middleName":"","lastName":"Venter","suffix":""},{"id":428802415,"identity":"395cdd8b-8ad9-40da-898d-39cf993d1767","order_by":7,"name":"Gcobisa Ndlangalavu","email":"","orcid":"","institution":"Stellenbosch University","correspondingAuthor":false,"prefix":"","firstName":"Gcobisa","middleName":"","lastName":"Ndlangalavu","suffix":""},{"id":428802416,"identity":"81db7b76-970c-436d-a2e2-6dd4bc9ed379","order_by":8,"name":"Byron W.P. Reeve","email":"","orcid":"","institution":"Stellenbosch University","correspondingAuthor":false,"prefix":"","firstName":"Byron","middleName":"W.P.","lastName":"Reeve","suffix":""},{"id":428802417,"identity":"b375de36-5c94-42dc-b64e-c9b160bfd435","order_by":9,"name":"Andrew M. Marino","email":"","orcid":"","institution":"University of Pennsylvania","correspondingAuthor":false,"prefix":"","firstName":"Andrew","middleName":"M.","lastName":"Marino","suffix":""},{"id":428802418,"identity":"8d94f2e3-1767-4dcc-9350-1fce18e5faa4","order_by":10,"name":"Tim J. Bull","email":"","orcid":"","institution":"City and St. George’s University of London, London","correspondingAuthor":false,"prefix":"","firstName":"Tim","middleName":"J.","lastName":"Bull","suffix":""},{"id":428802419,"identity":"2b1bca79-e825-48a2-82e1-74972629c580","order_by":11,"name":"Alaina M. Olson","email":"","orcid":"","institution":"University of Washington","correspondingAuthor":false,"prefix":"","firstName":"Alaina","middleName":"M.","lastName":"Olson","suffix":""},{"id":428802420,"identity":"83a26e4e-82ba-44d0-b231-b9dbc42fe322","order_by":12,"name":"Rachel Wood","email":"","orcid":"","institution":"University of Washington","correspondingAuthor":false,"prefix":"","firstName":"Rachel","middleName":"","lastName":"Wood","suffix":""},{"id":428802421,"identity":"2b7ed193-fdce-4809-89d6-0c95df777e95","order_by":13,"name":"Gerard A. Cangelosi","email":"","orcid":"","institution":"University of Washington","correspondingAuthor":false,"prefix":"","firstName":"Gerard","middleName":"A.","lastName":"Cangelosi","suffix":""},{"id":428802422,"identity":"fe7c9a2e-7679-4932-a092-3aa0ee846a52","order_by":14,"name":"Robin M. Warren","email":"","orcid":"","institution":"Stellenbosch University","correspondingAuthor":false,"prefix":"","firstName":"Robin","middleName":"M.","lastName":"Warren","suffix":""},{"id":428802423,"identity":"f3e1a235-a142-4705-a5b2-76bc06be39c6","order_by":15,"name":"Grant Theron","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAAAyElEQVRIiWNgGAWjYFACHjApR4IONogWY9K1JDYQrUM+vvfg44qKO+nzZ+QeYPhRw2DPT0iz4TG+ZMMzZ57lbriRl8DYc4yBWeIAIS1tPGaSjW2HczdI5Bgw8DYwsDEQocX8J1BLuvyMHAPGvw0MPPKEtMiz8ZgxArUkMNzIMWAG2iJhQEiLAVuOsWTDmcOGG868MTgsc0zCwJCgLc1nDD82VByWl2/PMXz4psbGXo6gLcgKgGwJAupBtjQQVjMKRsEoGAUjHQAAUv47y4fMmMEAAAAASUVORK5CYII=","orcid":"","institution":"Stellenbosch University","correspondingAuthor":true,"prefix":"","firstName":"Grant","middleName":"","lastName":"Theron","suffix":""}],"badges":[],"createdAt":"2025-03-14 10:56:53","currentVersionCode":1,"declarations":{"humanSubjects":true,"vertebrateSubjects":false,"conflictsOfInterestStatement":false,"humanSubjectEthicalGuidelines":true,"humanSubjectConsent":true,"humanSubjectClinicalTrial":false,"humanSubjectCaseReport":false,"vertebrateSubjectEthicalGuidelines":false},"doi":"10.21203/rs.3.rs-6225530/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-6225530/v1","draftVersion":[],"editorialEvents":[],"editorialNote":"","failedWorkflow":false,"files":[{"id":78746916,"identity":"cd88808b-183f-4e95-9987-6cae76b329e3","added_by":"auto","created_at":"2025-03-18 10:45:31","extension":"png","order_by":1,"title":"Figure 1","display":"","copyAsset":false,"role":"figure","size":82272,"visible":true,"origin":"","legend":"\u003cp\u003eFlow chart showing enrolment, specimen collection and processing, and tests in Cohorts A and B.\u003c/p\u003e","description":"","filename":"floatimage1.png","url":"https://assets-eu.researchsquare.com/files/rs-6225530/v1/91bb94b58347154415bd9d62.png"},{"id":78746917,"identity":"2ff33f19-90f2-4f31-8686-672253dd09d7","added_by":"auto","created_at":"2025-03-18 10:45:31","extension":"png","order_by":2,"title":"Figure 2","display":"","copyAsset":false,"role":"figure","size":1149171,"visible":true,"origin":"","legend":"\u003cp\u003eForest plot of Ultra sensitivity and specificity on different oral samples in Cohort A, overall and stratified by HIV. Oral washes, foam swabs, and flocked swabs has the highest sensitivity.\u003c/p\u003e","description":"","filename":"floatimage2.png","url":"https://assets-eu.researchsquare.com/files/rs-6225530/v1/d88e495c42634a3202aa9393.png"},{"id":78747368,"identity":"12bfb44e-02af-4eb5-a7ef-cab15fc6a522","added_by":"auto","created_at":"2025-03-18 10:53:31","extension":"png","order_by":3,"title":"Figure 3","display":"","copyAsset":false,"role":"figure","size":700403,"visible":true,"origin":"","legend":"\u003cp\u003eForest plot of Ultra and culture sensitivity and specificity on oral samples in Cohort B. OWs had the highest sensitivity of oral samples.\u003c/p\u003e","description":"","filename":"floatimage3.png","url":"https://assets-eu.researchsquare.com/files/rs-6225530/v1/1402ee5c7d4b2e3fb5434081.png"},{"id":78747364,"identity":"eafa8485-e716-4d99-9502-c3d46ecccf0c","added_by":"auto","created_at":"2025-03-18 10:53:31","extension":"png","order_by":4,"title":"Figure 4","display":"","copyAsset":false,"role":"figure","size":257763,"visible":true,"origin":"","legend":"\u003cp\u003eNumber of people with a positive result in Cohorts A and B. In both cohorts, more people were detected exclusively by OW Ultra than any other test. Each oral sampling method also detected people missed by other approaches. The first column includes people with an OW, whereas the second includes people with tongue swabs (in Cohort B, OWs and tongue swabs were in the same people but are disaggregated here for clarity).\u003c/p\u003e","description":"","filename":"floatimage4.png","url":"https://assets-eu.researchsquare.com/files/rs-6225530/v1/1854ba17ed3c7b7e41b043ce.png"},{"id":78748634,"identity":"60bf075f-894f-4702-825a-59f4d3c96d0f","added_by":"auto","created_at":"2025-03-18 11:09:33","extension":"pdf","order_by":0,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":3320680,"visible":true,"origin":"","legend":"","description":"","filename":"manuscript.pdf","url":"https://assets-eu.researchsquare.com/files/rs-6225530/v1/2fa3c251-0f35-4d9f-960a-f840ae5f89f2.pdf"},{"id":78746919,"identity":"3248699f-b171-42a2-8fe6-6ba41d5aba7a","added_by":"auto","created_at":"2025-03-18 10:45:31","extension":"docx","order_by":1,"title":"","display":"","copyAsset":false,"role":"supplement","size":101018,"visible":true,"origin":"","legend":"","description":"","filename":"OralUltraSupplementaryresultsLR13March.docx","url":"https://assets-eu.researchsquare.com/files/rs-6225530/v1/ff26b90041830c02566425fa.docx"}],"financialInterests":"The authors declare no competing interests.","formattedTitle":"\u003cp\u003e\u003cstrong\u003eOral washes and tongue swabs for Xpert MTB/RIF Ultra-based tuberculosis diagnosis in people with and without the ability to make sputum\u003c/strong\u003e\u003c/p\u003e","fulltext":[{"header":"Background","content":"\u003cp\u003eTuberculosis (TB), caused by \u003cem\u003eMycobacterium tuberculosis\u0026nbsp;\u003c/em\u003ecomplex (\u003cem\u003eMtb\u003c/em\u003e), is a global pandemic (1). Molecular tests like Xpert MTB/RIF Ultra (Ultra; Cepheid, Sunnyvale, USA) and Truenat MTB Plus (Molbio Diagnostics, Goa, India) were designed for sputum. Non-sputum tests are urgently needed to reach the millions of people with TB undiagnosed each year (2-4).\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eTongue swabs, which collect biofilm from the dorsum tongue, have data to support use with the Ultra and Truenat MTB Ultima (Ultima) tests: Ultra has sensitivities from 72-75% with high specificity, with similar performance described for Ultima (5). Furthermore, discrete choice studies show tongue swabs are more acceptable than sputum to some people despite potential sensitivity trade-offs (6). Importantly, a test with diminished sensitivity can increase diagnostic yield if the number of individuals tested is higher (7, 8), however, there are few empiric data in the context of oral samples (9). \u0026nbsp;\u003c/p\u003e\n\u003cp\u003eKey gaps remain regarding the diagnostic potential of oral samples, including samples other than tongue swabs. Early proof-of-concept studies used in-house rather than commercially-available tests or, as tongue swabs are likely most beneficial in people unable to expectorate, paradoxically pre-selected people based on their ability to naturally produce sputum (10-13). Furthermore, the type of tongue swab (flocked, foam) and processing method may influence performance (14). Equally important, data from people with risk factors for TB not preselected based on symptoms, who are less able to make sputum and increasingly targeted as part of facility active case finding strategies, are scarce. More data are urgently needed to inform global policy.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eData are also scarce regarding \u003cem\u003eMtb\u0026nbsp;\u003c/em\u003eculturability from tongue swabs. One study found 44-58% of swabs to be MGIT960 culture-positive in people with TB (15). Culturable \u003cem\u003eMtb\u0026nbsp;\u003c/em\u003efrom the swab could reinforce microbiological reference standards for diagnostic accuracy evaluations and provide material for drug susceptibility testing. Lastly, culture methods that do not involve harsh NALC-NaOH decontamination and are designed for paucibacillary samples may have added utility.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eTo address these knowledge gaps, we evaluated\u0026nbsp;different types of oral samples, swabs, and processing methods for TB diagnosis in a HIV high burden setting where people were offered sputum induction.\u0026nbsp;We hypothesized Ultra would detect \u003cem\u003eMtb\u003c/em\u003e in oral specimens with high concordance to sputum.\u0026nbsp;\u003cstrong\u003e\u003cbr\u003e\u0026nbsp;\u003c/strong\u003e\u003c/p\u003e"},{"header":"Methods and Materials","content":"\u003ch3\u003e\u003cu\u003eEthics statement\u003c/u\u003e\u003c/h3\u003e\n\u003cp\u003eThis work was approved by the Health Research Ethics Committee of Stellenbosch University (N14/10/136, M20/06/017, M20/06/018)\u0026nbsp;and City of Cape Town (10570)\u003cem\u003e.\u003c/em\u003e Written informed consent was obtained.\u003c/p\u003e\n\u003ch3\u003e\u003cu\u003eStudy Cohorts\u0026nbsp;\u003c/u\u003e\u003c/h3\u003e\n\u003cp\u003e\u003cem\u003eCohort A:\u003c/em\u003e Adults (\u0026ge;18 years) self-reporting with presumptive pulmonary TB symptoms meeting WHO criteria (at least one symptom for PLHIV, \u0026ge;2 for those without HIV) were recruited at clinics in Cape Town, South Africa.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e\u003cem\u003eCohort B:\u003c/em\u003e Antiretroviral therapy (ART) initiators (\u0026ge;18 years),\u0026nbsp;regardless of symptoms, were recruited\u0026nbsp;in Cape Town as previously described (16, 17).\u003c/p\u003e\n\u003ch3\u003e\u003cu\u003eDefinitions\u003c/u\u003e\u003c/h3\u003e\n\u003cp\u003e\u003cem\u003eCohort A:\u003c/em\u003e People were classified as having TB using an extended microbiological reference standard (eMRS) if sputum was Ultra-positive or had\u003cem\u003e\u0026nbsp;Mtb\u003c/em\u003e-positive culture growth. Those categorized as not having TB had no positive sputum Ultra or culture results at least one negative result. Those missing a culture or Ultra result were, if other results negative, classified as not having TB.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e\u003cem\u003eCohort B:\u003c/em\u003e The eMRS consisted of two sputum cultures. If at least one culture was \u003cem\u003eMtb-\u003c/em\u003epositive, participants were classified as having TB. Participants with negative culture(s) did not have TB.\u003c/p\u003e\n\u003ch3\u003e\u003cu\u003eSpecimen collection\u0026nbsp;\u003c/u\u003e\u003c/h3\u003e\n\u003cp\u003eAll oral samples were collected prior to sputum collection after\u0026nbsp;\u0026ge;30 min had passed since food or fluid was ingested and teeth brushed.\u0026nbsp;\u003cstrong\u003eFigure 1\u0026nbsp;\u003c/strong\u003esummarises specimen collection and testing.\u003c/p\u003e\n\u003cp\u003e\u003cem\u003eOral washes\u003c/em\u003e\u003c/p\u003e\n\u003cp\u003eEach person (n=40 Cohort A, n=156 Cohort B) received two vials of 20 mL sterile water. One vial was used to rinse the mouth and discarded. The second vial\u0026rsquo;s contents were swirled in the mouth for a few seconds and spat into the vial.\u0026nbsp;All samples were stored at\u0026nbsp;-20 \u0026deg;C until processed. In Cohort A, OWs were collected from separate people than tongue swabs. In Cohort B, OWs were collected in everyone who gave tongue swabs and collection was done immediately after swab collection.\u003c/p\u003e\n\u003cp\u003e\u003cem\u003eTongue swabs\u0026nbsp;\u003c/em\u003e\u003c/p\u003e\n\u003cp\u003eCohort A: One\u0026nbsp;dry flocked swab (FLOQSwabs code 520C; Copan Italia S.p.A., Brescia, Italy)\u0026nbsp;followed by a dry foam swab (Medline Industries, Northfield, USA) were consecutively collected from 851 and 550 people, respectively, by scraping the tongue dorsum for 10-15 s and placing the swab in a dry tube (550 people had both swab types). As a control for each person,\u0026nbsp;an air swab was collected in the same space as the participant by waving a foam swab (flocked prior to the start of foam swab sampling) in the air for 10-15 s immediately before sampling.\u003c/p\u003e\n\u003cp\u003eCohort B: One\u0026nbsp;flocked swab was\u0026nbsp;collected in 800 \u0026mu;L\u0026nbsp;Tris-EDTA (TE) buffer from 156 people. The last 122 were asked to provide an additional two flocked swabs subsequently pooled into a single tube with 800 \u0026mu;l TE buffer. A flocked air swab was done as for Cohort A. In both cohorts, samples were stored (-20 \u0026deg;C) until processed.\u003c/p\u003e\n\u003cp\u003e\u003cem\u003eBuccal swabs and periodontal brushes\u003c/em\u003e\u003c/p\u003e\n\u003cp\u003eIn Cohort B, paired buccal swabs and periodontal brushes (n=102) were collected (\u003cstrong\u003eSupplementary text pg. 2\u003c/strong\u003e) and stored (-20 \u0026deg;C) until processed.\u0026nbsp;Collection occurred prior to the start tongue swab and OW collection.\u003c/p\u003e\n\u003cp\u003e\u003cem\u003eSputum\u003c/em\u003e\u003c/p\u003e\n\u003cp\u003eCohort A participants were each asked to provide two sputa, while Cohort B participants were each asked to provide three sputa. Sputa were used for Ultra and MGIT960 culture [with 1% NALC-NaOH decontamination and MTBC confirmation using MTBDR\u003cem\u003eplus\u003c/em\u003e (Bruker-Hain Diagnostics, Nehren, Germany); one culture in Cohort A, two in Cohort B].\u0026nbsp;Sputum induction\u0026nbsp;was done (16, 18), however, it was only recorded whether people definitively required induction to make at least one \u0026ge;1 ml sputum in 584 Cohort A people and the 156 Cohort B people that gave tongue swabs.\u0026nbsp;\u003cstrong\u003eFigure 1\u0026nbsp;\u003c/strong\u003esummarises specimens collected\u003cstrong\u003e\u0026nbsp;\u003c/strong\u003eby Cohort.\u003c/p\u003e\n\u003ch3\u003e\u003cu\u003eSpecimen processing and testing\u003c/u\u003e\u003c/h3\u003e\n\u003cp\u003eAll processing was performed in a biosafety level 3 (BSL3) laboratory.\u003c/p\u003e\n\u003ch3\u003e\u003cem\u003eOral wash testing using Ultra (Cohorts A and B)\u003c/em\u003e\u003c/h3\u003e\n\u003cp\u003eIn Cohort A, OWs were concentrated, decontaminated, and processed with sample reagent (SR; Cepheid) per \u003cstrong\u003eSupplementary text pg. 2\u003c/strong\u003e.\u003cstrong\u003e\u0026nbsp;\u003c/strong\u003eDecontamination was not done in Cohort B.\u0026nbsp;\u003c/p\u003e\n\u003ch3\u003e\u003cem\u003eTongue swab testing using Ultra (Cohorts A and B)\u003c/em\u003e\u003c/h3\u003e\n\u003cp\u003eCohort A:\u0026nbsp;Swabs were removed from storage and placed into a heating block (100 \u0026deg;C, 10 min), after which TE buffer was added to 2 mL and the whole volume tested with Ultra (19) with\u0026nbsp;no SR.\u003c/p\u003e\n\u003cp\u003eCohort B: Single flocked swabs in TE buffer were removed from storage and immediately boiled (100\u0026deg;C, 10 min) followed by\u0026nbsp;SR\u0026nbsp;addition\u0026nbsp;(1.6 mL to 800 \u0026mu;L sample)\u0026nbsp;and Ultra (19). For the 122 people who also gave a double swab,\u0026nbsp;no heating was done, 2:1 SR added, and Ultra done \u003cstrong\u003e(Supplementary Figure 1).\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003e\u003cem\u003eAir swab controls in people who had tongue swabs (Cohorts A and B)\u003c/em\u003e\u003c/p\u003e\n\u003cp\u003eAir swabs was processed and tested with Ultra using the same procedure as tongue swabs every tenth patient. If a participant had a positive tongue swab their air swab was tested.\u003c/p\u003e\n\u003cp\u003e\u003cem\u003eBuccal swabs and periodontal brushes\u003c/em\u003e \u003cem\u003e(Cohort B)\u003c/em\u003e\u003c/p\u003e\n\u003cp\u003eStored samples were removed, processed, and tested with Ultra using the same procedure as tongue swabs in Cohort B.\u003c/p\u003e\n\u003ch3\u003e\u003cem\u003eTongue swab culture\u003c/em\u003e\u003cem\u003e\u0026nbsp;(Cohort B)\u0026nbsp;\u003c/em\u003e\u003c/h3\u003e\n\u003cp\u003eWe did different types of culture (MGIT960, TiKa (20), early bactericidal activity (EBA) (21); methodology in \u003cstrong\u003eSupplementary text pg. 2\u003c/strong\u003e) on single flocked tongue swabs (without heat inactivation). Speciation on positive growth was done using Ultra on a concentrated MGIT960 tube\u003cstrong\u003e\u0026nbsp;\u003c/strong\u003e(22).\u003c/p\u003e\n\u003ch3\u003e\u003cu\u003eAnalysis\u0026nbsp;\u003c/u\u003e\u003c/h3\u003e\n\u003cp\u003eMethods and reporting are per STARD guidelines (23). Diagnostic accuracy metrics were calculated using Excel (Microsoft, Redmond, USA) and compared using prtest (24) in STATA version 16.0 (StataCorp, Texas, USA).\u0026nbsp;The effect on sensitivity and specificity of Ultra trace results removed or reclassified to Ultra-negative in 2x2 tables was evaluated.\u0026nbsp;Continuous data were compared with GraphPad Prism version 8.0.1 (GraphPad Software, San Diego, USA), also used for linear regression and correlation analysis. Diagnostic yield (DYT, diagnostic yield in those tested; DYD, diagnostic yield in those diagnosed) was calculated as described\u0026nbsp;(8)\u0026nbsp;and defined in \u003cstrong\u003eSupplementary text pg. 3\u003c/strong\u003e. Morbidity score information (TBscoreII) was collected\u0026nbsp;(25).\u0026nbsp;Venn diagrams were made using Interactivenn\u0026nbsp;(26). In Cohort B, people were designated asymptomatic based on the WHO four symptom screen\u0026nbsp;(27).\u0026nbsp;P-values\u0026nbsp;\u0026le;0.05 were significant. Unsuccessful results are those not positive or negative by any test. We compared SPC C\u003csub\u003eT\u003c/sub\u003e values from Ultra to measure inhibition (lower SPC C\u003csub\u003eT\u003c/sub\u003es mean less inhibition)\u0026nbsp;(28).\u003c/p\u003e"},{"header":"Results","content":"\u003ch3\u003e\u003cu\u003eParticipant demographics\u003c/u\u003e\u003c/h3\u003e\n\u003cp\u003ePeople in Cohort A were, compared to Cohort B, more likely to be female, have higher morbidity, and more likely to have previous TB (\u003cstrong\u003eTable 1)\u003c/strong\u003e. In both cohorts, people with TB were more likely to be male and have higher morbidity.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e\u003cu\u003eDiagnostic accuracy of Ultra\u003c/u\u003e\u003c/p\u003e\n\u003cp\u003eData are summarised in \u003cstrong\u003eFigures 2 and 3\u003c/strong\u003e.\u0026nbsp;All air swabs were negative.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e\u003cem\u003eCohort A\u003c/em\u003e\u003c/p\u003e\n\u003cp\u003eOWs: No unsuccessful results occurred. Sensitivity was 80% (56, 94) and specificity 80% (56-94). Sensitivity was higher among those without HIV compared to PLHIV [94% (70-100) vs 25% (1-81); p=0.002] (\u003cstrong\u003eTable 2\u003c/strong\u003e).\u0026nbsp;Four false-positive results occurred (all trace semi-quantitation, two previous TB). \u0026nbsp;\u003c/p\u003e\n\u003cp\u003eTongue\u0026nbsp;swabs: 3% (25/851) of flocked swabs and 2% (13/550; p=0.611) of foam swabs had unsuccessful results (3 both, 22 flocked only, 10 foam only; mostly overpressure errors). Flocked swabs had lower sensitivity than foam swabs [59% (53-65) vs. 65% (58-72); p=0.001] with high specificities [94% (91-96) vs. 92% (89-95); p=0.100] (\u003cstrong\u003eTable 2\u003c/strong\u003e). Amongst\u0026nbsp;false-positive swabs,\u0026nbsp;26% (9/34) of flocked swabs and 39% (11/28) of foam swabs were from people programmatically empirically treated (no positive bacteriology at treatment start). Different trace recategorization strategies resulted in small sensitivity decreases and large specificity increases (\u003cstrong\u003eSupplementary Table 1\u003c/strong\u003e).\u003c/p\u003e\n\u003cp\u003e\u003cem\u003eCohort B\u003c/em\u003e\u003c/p\u003e\n\u003cp\u003eOWs: No unsuccessful results occurred. Sensitivity was 71% (42-92) and specificity 92% (86- 96). Of the 12 false positives, all were Ultra semi-quantitation category trace. 17% (2/12) were Ultra flocked tongue swab positive.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eBuccal swabs and periodontal brushes:\u0026nbsp;No unsuccessful results occurred.\u0026nbsp;Sensitivity was\u0026nbsp;7% (0-34) and\u0026nbsp;14% (2-43), respectively, with 98% (92-100) specificity for both (\u003cstrong\u003eSupplementary Table 2\u003c/strong\u003e).\u003c/p\u003e\n\u003cp\u003eTongue swabs: 1% (2/156) of single swabs, and 1% (1/122; p=0.711) double swabs generated unsuccessful results.\u0026nbsp;In head-to-head analyses, single vs. double swab sensitivity was 67% (30-93) and 22% (3-60; p=0.068) and specificity 82% (74-89) vs. 96% (91-99; p=0.009) (\u003cstrong\u003eTable 3\u003c/strong\u003e). Alternative evidence of TB occurred in 23% (6/26) of false-positive single flocked swabs [two sputum Ultra-positive, one\u0026nbsp;Alere Determine TB LAM Ag (Abbott, USA)-positive, three positive MGIT960 tongue swab culture] and 25% (1/4) of false-positive double swabs (sputum Ultra-positive). Different trace recategorization strategies resulted in small sensitivity decreases and large specificity increases (\u003cstrong\u003eSupplementary Table 3\u003c/strong\u003e).\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eAsymptomatic TB: 58% (150/258) of people were asymptomatic and, of these, 6% (9/150) had TB. 33% (3/9), 22% (2/9) and 0% (0/9) were positive using OW, single flocked swab and double flocked swabs, respectively.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e\u003cu\u003eDiagnostic accuracy of tongue swab culture in Cohort B\u003c/u\u003e\u003c/p\u003e\n\u003cp\u003eMGIT960 and TiKa had 64% (35-87) vs. 36% (13-65; p=0.131) sensitivity, while specificity was 88% (82-93) vs. 94% (89-98; p=0.060) (Supplementary \u003cstrong\u003eTable 4\u003c/strong\u003e). For comparison, sputum TiKa culture sensitivity and specificity were 79% (49-95; p=0.022 vs. TiKa on tongue swabs) and 97% (93-99; p=0.238), respectively. In the subset who underwent tongue swab EBA culture, sensitivity and specificity were 33% (4-78) and 95% (85-99).\u003c/p\u003e\n\u003ch3\u003e\u003cu\u003eYield of Ultra\u003c/u\u003e\u003c/h3\u003e\n\u003cp\u003e\u003cstrong\u003eFigure 4\u003c/strong\u003e shows people who tested positive by Ultra on different samples, as well as sputum culture. Ultra yield metrics are compared in \u003cstrong\u003eSupplementary\u0026nbsp;\u003c/strong\u003e\u003cstrong\u003eTable 5\u003c/strong\u003e. Similar patterns for DYD occurred as those described for DYT. We also calculate yields from tongue swab culture \u003cstrong\u003e(Supplementary text pg. 4)\u003c/strong\u003e, which were low.\u003c/p\u003e\n\u003cp\u003e\u003cem\u003eCohort A\u003c/em\u003e\u003c/p\u003e\n\u003cp\u003eOverall: OW DYT (50%, 20/50) was like that for sputum. Foam tongue swabs had a higher DYT (27%, 150/550) point estimate than flocked tongue swabs (23%, 131/150) but this did not reach significance. Sputum DYT was significantly higher [33% (180/550] than flocked and foam swabs.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eIf sputum induction were unavailable: No one with an OW had sputum induction information. In people with both tongue swab types, 5% (28/550) could not expectorate sputum. 25% (7/28) and 25% (7/28) were foam or flocked tongue swab positive. If tongue swabs were done in people who could not expectorate sputum, people bacteriologically diagnosed rapidly would increase from 175 (Ultra-positive on expectorated sputum) to 182 for flocked and foam swabs, [4% (2-6) increase].\u0026nbsp;\u003cem\u003e\u003cbr\u003e\u003c/em\u003e\u003c/p\u003e\n\u003cp\u003e\u003cem\u003eCohort B\u003c/em\u003e\u003c/p\u003e\n\u003cp\u003eOverall: Amongst people with all three sample types, DYT point estimates were highest for single flocked tongue swab compared to OWs and double tongue swabs [21% (26/122), 14% (17/122), 5% (6/122)]. Sputum DYT was\u0026nbsp;12% (5/122).\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eIf sputum induction were unavailable: In people who had all three sample types (OWs, single and double swabs), 35% (43/122) could not expectorate sputum, of which 16% (7/43), 26% (11/43) and 5% (2/43) were positive on each oral sample type. If oral sampling was done in people who could not expectorate sputum, bacteriological diagnoses would change from 8 (Ultra-positive on expectorated sputum) to 15, 19, and 10 for OWs, single swabs, and double swabs, respectively [+88% (82-93), +138% (130-145) and 25% (17-33) respectively. If induction were unavailable, sputum DYT decreased to 7% (8/122), lower than OWs [14% 17/122; p=0.057] and single flocked swabs [21% (26/122); p=0.0009].\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e\u003cu\u003eUltra inhibition\u003c/u\u003e\u003c/p\u003e\n\u003cp\u003eIn Cohort A, oral washes had less PCR inhibition than sputum, however, this did not occur in Cohort B. No other differences of note occurred across sample types \u003cstrong\u003e(Supplementary text pg. 4).\u003c/strong\u003e\u003cem\u003e\u003cbr\u003e\u0026nbsp;\u003c/em\u003e\u003c/p\u003e"},{"header":"Discussion","content":"\u003cp\u003eOur key findings are: 1) OW Ultra had the highest sensitivity amongst the methods tested (71-80%), 2) Ultra on foam swabs had higher sensitivity than flocked swabs (65% vs. 59%; Cohort A), 3) Ultra on oral samples swabs diagnosed TB in many people who could not naturally expectorate, permitting yield to exceed that of sputum-based testing in Cohort B where sputum scarcity was more common, and 4) other approaches (Ultra on double tongue swabs, buccal swabs, and periodontal brushes) were suboptimal. These data suggest oral sampling, especially oral washes and foam swabs, can improve the diagnosis of TB, especially when sputum scarcity is accounted for.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eOW Ultra had a higher sensitivity point estimate than other approaches. To our knowledge, we report the first study using OW Ultra for TB diagnosis. Previous studies (29, 30) used in-house PCR methods applied to OW reported sensitivities of 77% and 88%, broadly our findings. We suggest OWs are included as a comparator in research on tongue swabs for TB diagnosis going forward as, unlike swabs, they require less processing.\u003c/p\u003e\n\u003cp\u003eWe observed a small sensitivity increase with foam swabs compared to flocked swabs, likely attributable to the larger amount of biomass bound to the foam swab (15, 31).\u0026nbsp;Foam swabs have an added advantage in that they are relatively\u0026nbsp;cheap (0.14 USD per swab) compared to flocked swabs. On a practical note, however, participants report foam swabs make their tongues feel \u0026ldquo;dry\u0026rdquo;. Furthermore, the lack of a breakpoint means swab heads need to be cut off, however, this is addressable. The magnitude of the sensitivity improvement from foam swabs may, vs. flocked swabs, be even greater in people with earlier stage disease.\u003c/p\u003e\n\u003cp\u003eWe showed Ultra on swabs can detect TB in people who cannot make. As TB testing programmatically expands, including to people with early stage potentially subclinical TB, the proportion of people with sputum scarce TB that require testing will increase. Thus, our study addresses a key gap: most studies on oral samples for TB diagnosis have either not recruited people with sputum scarce TB or have offered induction and not been able to disaggregate performance in people who, without induction, cannot expectorate. This is important because, even if tongue swabs perform well in sputum expectorators, it is hard to justify not testing sputum if it is available. Our data are thus novel in that they give performance data in a type of person most likely to benefit from non-sputum tests. For example, in Cohort B, the number of people with a positive Ultra result would at least double with the use of oral sampling in sputum scarce.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eWe compared a single flocked swab with heat lysis, which had high sensitivity in earlier work using contrived samples\u0026nbsp;(32),\u0026nbsp;to a double flocked swab without heat lysis, which had lower sensitivity. This suggests heating is critical to release DNA. It remains to be seen if heating with a double swab improves sensitivity further, however, given our results from foam swabs, it is likely that steps that input material would be beneficial, providing heat lysis occurs. In addition to the double swab method, we evaluated other specimen types (buccal swabs, periodontal brushes), as well as different tongue swabs culture methods, however, neither were promising.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eOur study has strengths and limitations. Recruitment was programmatic in nature in that it included people with symptoms and non-symptom-based risk factors for which South African guidelines (33, 34) require molecular testing, even if they were not yet symptomatic and/or could not expectorate sputum. More data are needed in these groups where oral sample testing is likely to be most impactful (7).\u0026nbsp;Different processing methods, such as those which maximise time to recover DNA buccal swabs and periodontal brushes were in buffer for short periods) should be further explored alongside OWs, which notably does not require heating. Lastly, the inclusion of people who cannot make sputum naturally and crucially the availability of induction information are strengths.\u0026nbsp;\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eIn conclusion, Ultra on oral samples \u0026ndash; especially foam tongue swabs and OW \u0026ndash; is sensitive and highly specific and can significantly increase the overall number of people with a rapid positive bacteriological result when applied to people who cannot naturally make sputum. \u0026nbsp;\u003cbr\u003e\u0026nbsp;\u003c/p\u003e"},{"header":"Declarations","content":"\u003cp\u003e\u003cu\u003eFunding:\u003c/u\u003e South African Medical Research Council, National Research Foundation\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eAcknowledgements\u003c/p\u003e\n\u003cp\u003eThe authors thank the National Health Laboratory Services, Cape Town, South Africa,\u0026nbsp;members of the Clinical Mycobacteriology \u0026amp; Epidemiology (CLIME) research group, and study participants.\u0026nbsp;Research reported in this publication was supported by the South African Medical Research Council (SAMRC),\u0026nbsp;the National Research Foundation (NRF) and the postgraduate bursary department of SU are hereby also acknowledged for funding this project. LR acknowledges funding from the SAMRC. The work reported herein was made possible through funding by the SAMRC through its Division of Research Capacity Development under the SAMRC Internship Scholarship Programme. The content hereof is the sole responsibility of the authors and does not necessarily represent the official views of the SAMRC. GT reports funding from the EDCTP2 programme supported by the EU (RIA2018D-2509, PreFIT; RIA2018D-2493, SeroSelectTB; RIA2020I-3305, CAGE-TB) and the National Institutes of Health (D43TW010350, U01AI152087, U54EB027049, and R01AI136894).\u0026nbsp;Cepheid donated cartridges used in the study. They did not, however, have a role in study design or result interpretation.\u003c/p\u003e\n"},{"header":"References","content":"\u003col\u003e\n\u003cli\u003eWorld Health Organization. Global Tuberculosis Report 2023. World Health Organization, Geneva, Switzerland. 2023.\u003c/li\u003e\n\u003cli\u003eWorld Health Organization. Target product profile for tuberculosis diagnosis and detection of drug resistance. World Health Organization, Geneva, Switzerland. 2024.\u003c/li\u003e\n\u003cli\u003eNathavitharana RR, Garcia-Basteiro AL, Ruhwald M, Cobelens F, Theron G. Reimagining the status quo: How close are we to rapid sputum-free tuberculosis diagnostics for all? EBioMedicine. 2022;78.\u003c/li\u003e\n\u003cli\u003eAbdulgader SM, Okunola AO, Ndlangalavu G, Reeve BW, Allwood BW, Koegelenberg CF, et al. Diagnosing tuberculosis: what do new technologies allow us to (not) do? Respiration. 2022;101(9):797-813.\u003c/li\u003e\n\u003cli\u003eWood RC, Luabeya AK, Dragovich RB, Olson AM, Lochner KA, Weigel KM, et al. Diagnostic accuracy of tongue swab testing on two automated tuberculosis diagnostic platforms, Cepheid Xpert MTB/RIF Ultra and Molbio Truenat MTB Ultima. Journal of Clinical Microbiology. 2024;62(4):e00019-24.\u003c/li\u003e\n\u003cli\u003eShah K, Nalugwa T, Marcelo D, Nakawunde R, Trinh T, Shankar D, et al. A Multi-Country Evaluation of Patient Preferences for Tuberculosis Diagnostics: A Discrete Choice Experiment to Inform WHO Target Product Profiles. medRxiv. 2024:2024.06.19.24309124.\u003c/li\u003e\n\u003cli\u003ede Nooy A, Ockhuisen T, Korobitsyn A, Khan SA, Ruhwald M, Ismail N, et al. Trade-offs between clinical performance and test accessibility in tuberculosis diagnosis: a multi-country modelling approach for target product profile development. The Lancet Global Health. 2024;12(7):e1139-e48.\u003c/li\u003e\n\u003cli\u003eBroger T, Marx FM, Theron G, Marais BJ, Nicol MP, Kerkhoff AD, et al. Diagnostic yield as an important metric for the evaluation of novel tuberculosis tests: rationale and guidance for future research. The Lancet Global Health. 2024;12(7):e1184-e91.\u003c/li\u003e\n\u003cli\u003eBezuidenhout C, Long L, Nichols B, Meyer-Rath G, Fox MP, Olifant S, et al. USING SPUTUM AND TONGUE SWAB SPECIMENS FOR IN-HOME POINT-OF-CARE TARGETED UNIVERSAL TESTING FOR TB OF HOUSEHOLD CONTACTS: AN ACCEPTABILITY AND FEASIBILITY ANALYSIS. medRxiv. 2024.\u003c/li\u003e\n\u003cli\u003eValinetz ED, Cangelosi GA. A Look Inside: Oral Sampling for Detection of Non-Oral Infectious Diseases. Journal of Clinical Microbiology. 2021:JCM. 02360-20.\u003c/li\u003e\n\u003cli\u003eLuabeya AK, Wood RC, Shenje J, Filander E, Ontong C, Mabwe S, et al. Noninvasive detection of tuberculosis by oral swab analysis. Journal of clinical microbiology. 2019;57(3):e01847-18.\u003c/li\u003e\n\u003cli\u003eNicol MP, Wood RC, Workman L, Prins M, Whitman C, Ghebrekristos Y, et al. Microbiological diagnosis of pulmonary tuberculosis in children by oral swab polymerase chain reaction. Scientific reports. 2019;9(1):10789.\u003c/li\u003e\n\u003cli\u003eWood RC, Luabeya AK, Weigel KM, Wilbur AK, Jones-Engel L, Hatherill M, et al. Detection of Mycobacterium tuberculosis DNA on the oral mucosa of tuberculosis patients. Scientific reports. 2015;5(1):1-5.\u003c/li\u003e\n\u003cli\u003eChurch EC, Steingart KR, Cangelosi GA, Ruhwald M, Kohli M, Shapiro AE. Oral swabs with a rapid molecular diagnostic test for pulmonary tuberculosis in adults and children: a systematic review. The Lancet Global Health. 2024;12(1):e45-e54.\u003c/li\u003e\n\u003cli\u003eWood RC, Andama A, Hermansky G, Burkot S, Asege L, Job M, et al. Characterization of oral swab samples for diagnosis of pulmonary tuberculosis. PloS one. 2021;16(5):e0251422.\u003c/li\u003e\n\u003cli\u003eReeve BW, Ndlangalavu G, Mishra H, Palmer Z, Tshivhula H, Rockman L, et al. Point-of-care C-reactive protein and Xpert MTB/RIF Ultra for tuberculosis screening and diagnosis in unselected antiretroviral therapy initiators: a prospective, cross-sectional, diagnostic accuracy study. The Lancet Global Health. 2024;12(5):e793-e803.\u003c/li\u003e\n\u003cli\u003eMann T, Gupta RK, Reeve BW, Ndlangalavu G, Chandran A, Krishna AP, et al. Blood RNA biomarkers for tuberculosis screening in people living with HIV before antiretroviral therapy initiation: a diagnostic accuracy study. The Lancet Global Health. 2024;12(5):e783-e92.\u003c/li\u003e\n\u003cli\u003eMishra H, Reeve BW, Palmer Z, Caldwell J, Dolby T, Naidoo CC, et al. Xpert MTB/RIF Ultra and Xpert MTB/RIF for diagnosis of tuberculosis in an HIV-endemic setting with a high burden of previous tuberculosis: a two-cohort diagnostic accuracy study. The Lancet Respiratory Medicine. 2020;8(4):368-82.\u003c/li\u003e\n\u003cli\u003eCepheid. Xpert MTB/RIF Ultra product brochure. 2017.\u003c/li\u003e\n\u003cli\u003eGoosen WJ, Kleynhans L, Kerr TJ, van Helden PD, Buss P, Warren RM, et al. Improved detection of Mycobacterium tuberculosis and M. bovis in African wildlife samples using cationic peptide decontamination and mycobacterial culture supplementation. Journal of Veterinary Diagnostic Investigation. 2022;34(1):61-7.\u003c/li\u003e\n\u003cli\u003eDiacon AH, Donald PR. The early bactericidal activity of antituberculosis drugs. Expert review of anti-infective therapy. 2014;12(2):223-37.\u003c/li\u003e\n\u003cli\u003eGhebrekristos YT, Beylis N, Centner CM, Venter R, Derendinger B, Tshivhula H, et al. Xpert MTB/RIF Ultra on contaminated liquid cultures for tuberculosis and rifampicin-resistance detection: a diagnostic accuracy evaluation. The Lancet Microbe. 2023;4(10):e822-e9.\u003c/li\u003e\n\u003cli\u003eCohen JF, Korevaar DA, Altman DG, Bruns DE, Gatsonis CA, Hooft L, et al. STARD 2015 guidelines for reporting diagnostic accuracy studies: explanation and elaboration. BMJ open. 2016;6(11).\u003c/li\u003e\n\u003cli\u003eStataCorp L. prtest-Tests of proportions [cited 2023 Oct 22]. Available from: https://www.stata.com/manuals/rprtest.pdf.\u003c/li\u003e\n\u003cli\u003eRudolf F, Lemvik G, Abate E, Verkuilen J, Sch\u0026ouml;n T, Gomes VF, et al. TBscore II: refining and validating a simple clinical score for treatment monitoring of patients with pulmonary tuberculosis. Scandinavian journal of infectious diseases. 2013;45(11):825-36.\u003c/li\u003e\n\u003cli\u003eHeberle H, Meirelles GV, da Silva FR, Telles GP, Minghim R. InteractiVenn: a web-based tool for the analysis of sets through Venn diagrams. BMC bioinformatics. 2015;16(1):1-7.\u003c/li\u003e\n\u003cli\u003eWorld Health Organization. WHO consolidated guidelines on tuberculosis. Module 2: screening-systematic screening for tuberculosis disease. 2021.\u003c/li\u003e\n\u003cli\u003eBlakemore R, Nabeta P, Davidow AL, Vadwai V, Tahirli R, Munsamy V, et al. A multisite assessment of the quantitative capabilities of the Xpert MTB/RIF assay. American journal of respiratory and critical care medicine. 2011;184(9):1076-84.\u003c/li\u003e\n\u003cli\u003eDavis JL, Huang L, Kovacs JA, Masur H, Murray P, Havlir DV, et al. Polymerase chain reaction of secA1 on sputum or oral wash samples for the diagnosis of pulmonary tuberculosis. Clinical infectious diseases. 2009;48(6):725-32.\u003c/li\u003e\n\u003cli\u003eEvans D, Goyal M, Taylor I, Shaw R. Identification of M. tuberculosis ribosomal RNA in mouthwash samples from patients with tuberculosis. Respiratory medicine. 1994;88(9):687-91.\u003c/li\u003e\n\u003cli\u003eOlson AM, Wood RC, Weigel KM, Yan AJ, Lochner KA, Dragovich RB, et al. High-sensitivity detection of Mycobacterium tuberculosis DNA in tongue swab samples. Journal of Clinical Microbiology. 2025;63(2):e01140-24.\u003c/li\u003e\n\u003cli\u003eAndama A, Reza T, Jaganath D, Mulondo J, Semitala F, Wobudeya E, et al. Accuracy of Oral Swab Testing Using Xpert MTB-RIF Ultra for Tuberculosis Diagnosis. C106 FIRE TB LIVES ON: American Thoracic Society; 2022. p. A4928-A.\u003c/li\u003e\n\u003cli\u003eMartinson NA, Nonyane BA, Genade LP, Berhanu RH, Naidoo P, Brey Z, et al. Evaluating systematic targeted universal testing for tuberculosis in primary care clinics of South Africa: A cluster-randomized trial (The TUTT Trial). Plos Medicine. 2023;20(5):e1004237.\u003c/li\u003e\n\u003cli\u003eDepartment of Health SA. TB Screening and Testing Standard Operating Procedure,. 2022.\u003c/li\u003e\n\u003c/ol\u003e"},{"header":"Tables","content":"\u003cp\u003e\u003cstrong\u003eTable 1:\u0026nbsp;\u003c/strong\u003eDemographic and clinical characteristics. Data are n/N (%) or median (IQR).\u0026nbsp;\u003c/p\u003e\n\u003ctable border=\"1\" cellspacing=\"0\" cellpadding=\"0\"\u003e\n \u003ctbody\u003e\n \u003ctr\u003e\n \u003ctd rowspan=\"2\"\u003e\n \u003cp\u003e\u003cstrong\u003e\u0026nbsp;\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"3\"\u003e\n \u003cp\u003e\u003cstrong\u003eCohort A\u0026nbsp;\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"3\"\u003e\n \u003cp\u003e\u003cstrong\u003eCohort B\u0026nbsp;\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp\u003eOverall\u003c/p\u003e\n \u003cp\u003e\u0026nbsp;(n=891)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003eDefinite TB\u003c/p\u003e\n \u003cp\u003e\u0026nbsp;(n=315)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003eNon-TB\u003c/p\u003e\n \u003cp\u003e\u0026nbsp;(n=576)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003eOverall\u003c/p\u003e\n \u003cp\u003e\u0026nbsp;(n=258)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003eDefinite TB\u003c/p\u003e\n \u003cp\u003e\u0026nbsp;(n=28)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003eNon-TB\u003c/p\u003e\n \u003cp\u003e\u0026nbsp;(n=230)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd colspan=\"7\"\u003e\n \u003cp\u003e\u003cstrong\u003eDemographic\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp\u003eAge (years)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e38\u0026nbsp;\u003c/p\u003e\n \u003cp\u003e(30-48)\u003c/p\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e38\u0026nbsp;\u003c/p\u003e\n \u003cp\u003e(30-47)\u003c/p\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e38\u0026nbsp;\u003c/p\u003e\n \u003cp\u003e(30 - 48)\u003c/p\u003e\n \u003cp\u003ep=0.364*\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e37\u003c/p\u003e\n \u003cp\u003e\u0026nbsp;(31-44)\u003c/p\u003e\n \u003cp\u003ep=0.956\u003csup\u003e\u0026euro;\u003c/sup\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e36\u0026nbsp;\u003c/p\u003e\n \u003cp\u003e(31-44)\u003c/p\u003e\n \u003cp\u003ep=0.378\u003csup\u003e\u0026euro;\u003c/sup\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e39\u0026nbsp;\u003c/p\u003e\n \u003cp\u003e(31 -43)\u003c/p\u003e\n \u003cp\u003ep=0.352*\u003c/p\u003e\n \u003cp\u003e\u003cstrong\u003ep=0.002\u003c/strong\u003e\u003cstrong\u003e\u003csup\u003e\u0026euro;\u003c/sup\u003e\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp\u003eFemale\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e445/891\u0026nbsp;\u003c/p\u003e\n \u003cp\u003e(50)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e131/315\u003c/p\u003e\n \u003cp\u003e\u0026nbsp;(42)\u003c/p\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e314/576\u0026nbsp;\u003c/p\u003e\n \u003cp\u003e(55)\u003c/p\u003e\n \u003cp\u003e\u003cstrong\u003ep=0.002\u003c/strong\u003e\u003cstrong\u003e*\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e156/258\u003c/p\u003e\n \u003cp\u003e\u0026nbsp;(60)\u003c/p\u003e\n \u003cp\u003e\u003cstrong\u003ep=0.003\u003c/strong\u003e\u003cstrong\u003e\u003csup\u003e\u0026euro;\u003c/sup\u003e\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e11/28\u003c/p\u003e\n \u003cp\u003e\u0026nbsp;(39)\u003c/p\u003e\n \u003cp\u003ep=0.813\u003csup\u003e\u0026euro;\u003c/sup\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e145/230\u0026nbsp;\u003c/p\u003e\n \u003cp\u003e(63)\u003c/p\u003e\n \u003cp\u003e\u003cstrong\u003ep=0.015\u003c/strong\u003e\u003cstrong\u003e*\u003c/strong\u003e\u003c/p\u003e\n \u003cp\u003e\u003cstrong\u003ep=0.027\u003c/strong\u003e\u003cstrong\u003e\u003csup\u003e\u0026euro;\u003c/sup\u003e\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd colspan=\"7\"\u003e\n \u003cp\u003e\u003cstrong\u003eClinical\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp\u003eHIV-positive\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e355/891\u003c/p\u003e\n \u003cp\u003e\u0026nbsp;(40)\u003c/p\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e136/315\u003c/p\u003e\n \u003cp\u003e\u0026nbsp;(44)\u003c/p\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e219/576\u0026nbsp;\u003c/p\u003e\n \u003cp\u003e(38)\u003c/p\u003e\n \u003cp\u003ep=0.133*\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e258/258\u0026nbsp;\u003c/p\u003e\n \u003cp\u003e(100)\u003c/p\u003e\n \u003cp\u003e\u003cstrong\u003ep\u0026lt;0.0001\u003c/strong\u003e\u003cstrong\u003e\u003csup\u003e\u0026euro;\u003c/sup\u003e\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e28/28\u003c/p\u003e\n \u003cp\u003e\u0026nbsp;(100)\u003c/p\u003e\n \u003cp\u003e\u003cstrong\u003ep\u0026lt;0.0001\u003c/strong\u003e\u003cstrong\u003e\u003csup\u003e\u0026euro;\u003c/sup\u003e\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e230/230\u003c/p\u003e\n \u003cp\u003e\u0026nbsp;(100)\u003c/p\u003e\n \u003cp\u003e\u003cstrong\u003ep\u0026lt;0.0001\u003c/strong\u003e\u003cstrong\u003e\u003csup\u003e\u0026euro;\u003c/sup\u003e\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp\u003eCD4 count (cells/\u0026mu;L)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e305\u0026nbsp;\u003c/p\u003e\n \u003cp\u003e(269-903)\u003c/p\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e582\u0026nbsp;\u003c/p\u003e\n \u003cp\u003e(250-945)\u003c/p\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e305\u003c/p\u003e\n \u003cp\u003e\u0026nbsp;(290-605)\u003c/p\u003e\n \u003cp\u003ep=0.587*\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e293\u0026nbsp;\u003c/p\u003e\n \u003cp\u003e(143-434)\u003c/p\u003e\n \u003cp\u003e\u003cstrong\u003ep\u0026lt;0.0001\u003c/strong\u003e\u003cstrong\u003e\u003csup\u003e\u0026euro;\u003c/sup\u003e\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e59\u003c/p\u003e\n \u003cp\u003e\u0026nbsp;(22-135)\u003c/p\u003e\n \u003cp\u003e\u003cstrong\u003ep\u0026lt;0.0001\u003c/strong\u003e\u003cstrong\u003e\u003csup\u003e\u0026euro;\u003c/sup\u003e\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e299\u0026nbsp;\u003c/p\u003e\n \u003cp\u003e(171 - 438)\u003c/p\u003e\n \u003cp\u003e\u003cstrong\u003ep=0.013\u003c/strong\u003e\u003cstrong\u003e*\u003c/strong\u003e\u003c/p\u003e\n \u003cp\u003e\u003cstrong\u003ep\u0026lt;0.0001\u003c/strong\u003e\u003cstrong\u003e\u003csup\u003e\u0026euro;\u003c/sup\u003e\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp\u003eTBScoreII\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e3\u0026nbsp;\u003c/p\u003e\n \u003cp\u003e(2-3)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e3\u0026nbsp;\u003c/p\u003e\n \u003cp\u003e(3-4)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e3\u003c/p\u003e\n \u003cp\u003e\u0026nbsp;(2-3)\u003c/p\u003e\n \u003cp\u003e\u003cstrong\u003ep\u0026lt;0.0001\u003c/strong\u003e\u003cstrong\u003e*\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e2\u0026nbsp;\u003c/p\u003e\n \u003cp\u003e(0-2)\u003c/p\u003e\n \u003cp\u003e\u003cstrong\u003ep\u0026lt;0.0001\u003c/strong\u003e\u003cstrong\u003e\u003csup\u003e\u0026euro;\u003c/sup\u003e\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e3\u0026nbsp;\u003c/p\u003e\n \u003cp\u003e(2-4)\u003c/p\u003e\n \u003cp\u003e\u003cstrong\u003ep=0.001\u003c/strong\u003e\u003cstrong\u003e\u003csup\u003e\u0026euro;\u003c/sup\u003e\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e1\u0026nbsp;\u003c/p\u003e\n \u003cp\u003e(0-2)\u003c/p\u003e\n \u003cp\u003e\u003cstrong\u003ep=0.0006\u003c/strong\u003e\u003cstrong\u003e*\u003c/strong\u003e\u003c/p\u003e\n \u003cp\u003e\u003cstrong\u003ep\u0026lt;0.0001\u003c/strong\u003e\u003cstrong\u003e\u003csup\u003e\u0026euro;\u003c/sup\u003e\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp\u003ePrevious TB\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e398/891\u003c/p\u003e\n \u003cp\u003e\u0026nbsp;(45)\u003c/p\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e114/315\u003c/p\u003e\n \u003cp\u003e\u0026nbsp;(36)\u003c/p\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e284/576\u0026nbsp;\u003c/p\u003e\n \u003cp\u003e(49)\u003c/p\u003e\n \u003cp\u003e\u003cstrong\u003ep=0.0002\u003c/strong\u003e\u003cstrong\u003e*\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e26/258\u0026nbsp;\u003c/p\u003e\n \u003cp\u003e(10)\u003c/p\u003e\n \u003cp\u003e\u003cstrong\u003ep\u0026lt;0.0001\u003c/strong\u003e\u003cstrong\u003e\u003csup\u003e\u0026euro;\u003c/sup\u003e\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e2/28\u0026nbsp;\u003c/p\u003e\n \u003cp\u003e(7)\u003c/p\u003e\n \u003cp\u003e\u003cstrong\u003ep=0.002\u003c/strong\u003e\u003cstrong\u003e\u003csup\u003e\u0026euro;\u003c/sup\u003e\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e24/230\u003c/p\u003e\n \u003cp\u003e\u0026nbsp;(10)\u003c/p\u003e\n \u003cp\u003ep=0.585*\u003c/p\u003e\n \u003cp\u003e\u003cstrong\u003ep\u0026lt;0.0001\u003c/strong\u003e\u003cstrong\u003e\u003csup\u003e\u0026euro;\u003c/sup\u003e\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003c/tbody\u003e\n\u003c/table\u003e\n\u003cp\u003eWithin column p-values: *Definite TB vs. Non-TB, \u003csup\u003e\u0026euro;\u003c/sup\u003eAcross cohorts for people of the same TB status\u003c/p\u003e\n\u003cp\u003eAbbreviation: CD4, cluster of differentiation 4; HIV, human immunodeficiency virus; IQR, interquartile range\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eTable 2:\u0026nbsp;\u003c/strong\u003eDiagnostic accuracy of Ultra on OW or tongue swabs compared to an eMRS for the detection of TB stratified by HIV status in Cohort A. Data are %, (95% CI), and n/N.\u003c/p\u003e\n\u003ctable border=\"1\" cellspacing=\"0\" cellpadding=\"0\"\u003e\n \u003ctbody\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 84px;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"4\" valign=\"top\" style=\"width: 284px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eOverall\u003c/strong\u003e\u003cstrong\u003e\u003csup\u003e\u0026beta;\u003c/sup\u003e\u003c/strong\u003e\u003c/p\u003e\n \u003cp\u003e\u003cstrong\u003e(n=891)\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"4\" valign=\"top\" style=\"width: 254px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eHIV-negative\u003c/strong\u003e\u003c/p\u003e\n \u003cp\u003e\u003cstrong\u003e(n=526)\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"4\" valign=\"top\" style=\"width: 260px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eHIV-positive\u003c/strong\u003e\u003c/p\u003e\n \u003cp\u003e\u003cstrong\u003e(n=355)\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 84px;\"\u003e\n \u003cp\u003e\u003cstrong\u003e\u0026nbsp;\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 68px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eSensitivity\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 68px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eSpecificity\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 60px;\"\u003e\n \u003cp\u003e\u003cstrong\u003ePPV\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 89px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eNPV\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 68px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eSensitivity\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 68px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eSpecificity\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 58px;\"\u003e\n \u003cp\u003e\u003cstrong\u003ePPV\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 60px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eNPV\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 68px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eSensitivity\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 68px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eSpecificity\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 60px;\"\u003e\n \u003cp\u003e\u003cstrong\u003ePPV\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 64px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eNPV\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 84px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eOral wash\u0026nbsp;\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 68px;\"\u003e\n \u003cp\u003e80\u003c/p\u003e\n \u003cp\u003e\u0026nbsp;(56, 94)\u003c/p\u003e\n \u003cp\u003e16/20\u003c/p\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 68px;\"\u003e\n \u003cp\u003e80\u003c/p\u003e\n \u003cp\u003e\u0026nbsp;(56, 94)\u003c/p\u003e\n \u003cp\u003e16/20\u003c/p\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 60px;\"\u003e\n \u003cp\u003e80\u003c/p\u003e\n \u003cp\u003e\u0026nbsp;(56, 94)\u003c/p\u003e\n \u003cp\u003e16/20\u003c/p\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 89px;\"\u003e\n \u003cp\u003e80\u0026nbsp;\u003c/p\u003e\n \u003cp\u003e(56, 94)\u003c/p\u003e\n \u003cp\u003e16/20\u003c/p\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 68px;\"\u003e\n \u003cp\u003e94\u0026nbsp;\u003c/p\u003e\n \u003cp\u003e(70, 100)\u003c/p\u003e\n \u003cp\u003e15/16\u003c/p\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 68px;\"\u003e\n \u003cp\u003e93\u003c/p\u003e\n \u003cp\u003e\u0026nbsp;(57, 96)\u003c/p\u003e\n \u003cp\u003e14/17\u003c/p\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 58px;\"\u003e\n \u003cp\u003e83\u0026nbsp;\u003c/p\u003e\n \u003cp\u003e(59, 96)\u003c/p\u003e\n \u003cp\u003e15/18\u003c/p\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 60px;\"\u003e\n \u003cp\u003e93\u0026nbsp;\u003c/p\u003e\n \u003cp\u003e(68, 100)\u003c/p\u003e\n \u003cp\u003e14/15\u003c/p\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 68px;\"\u003e\n \u003cp\u003e25\u003c/p\u003e\n \u003cp\u003e\u0026nbsp;(1, 81)\u003c/p\u003e\n \u003cp\u003e1/4\u003c/p\u003e\n \u003cp\u003e\u003cstrong\u003ep=0.002\u003c/strong\u003e*\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 68px;\"\u003e\n \u003cp\u003e67\u0026nbsp;\u003c/p\u003e\n \u003cp\u003e(9, 99)\u003c/p\u003e\n \u003cp\u003e2/3\u003c/p\u003e\n \u003cp\u003ep=0.531*\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 60px;\"\u003e\n \u003cp\u003e50\u0026nbsp;\u003c/p\u003e\n \u003cp\u003e(1, 99)\u003c/p\u003e\n \u003cp\u003e1/2\u003c/p\u003e\n \u003cp\u003ep=0.264*\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 64px;\"\u003e\n \u003cp\u003e40\u0026nbsp;\u003c/p\u003e\n \u003cp\u003e(5, 85)\u003c/p\u003e\n \u003cp\u003e2/5\u003c/p\u003e\n \u003cp\u003e\u003cstrong\u003ep=0.010*\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 84px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eFlocked swab\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 68px;\"\u003e\n \u003cp\u003e59\u0026nbsp;\u003c/p\u003e\n \u003cp\u003e(53, 65)\u003c/p\u003e\n \u003cp\u003e164/277\u003c/p\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 68px;\"\u003e\n \u003cp\u003e94\u003c/p\u003e\n \u003cp\u003e\u0026nbsp;(91, 96)\u003c/p\u003e\n \u003cp\u003e515/549\u003c/p\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 60px;\"\u003e\n \u003cp\u003e83\u003c/p\u003e\n \u003cp\u003e\u0026nbsp;(77, 88)\u003c/p\u003e\n \u003cp\u003e164/198\u003c/p\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 89px;\"\u003e\n \u003cp\u003e82\u0026nbsp;\u003c/p\u003e\n \u003cp\u003e(79, 85)\u003c/p\u003e\n \u003cp\u003e515/628\u003c/p\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 68px;\"\u003e\n \u003cp\u003e56\u003c/p\u003e\n \u003cp\u003e\u0026nbsp;(48, 64)\u003c/p\u003e\n \u003cp\u003e\u0026nbsp; 84/150\u003c/p\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 68px;\"\u003e\n \u003cp\u003e82\u0026nbsp;\u003c/p\u003e\n \u003cp\u003e(90, 96)\u003c/p\u003e\n \u003cp\u003e\u0026nbsp;307/329\u003c/p\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 58px;\"\u003e\n \u003cp\u003e79\u003c/p\u003e\n \u003cp\u003e\u0026nbsp;(70, 87)\u003c/p\u003e\n \u003cp\u003e\u0026nbsp;84/106\u003c/p\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 60px;\"\u003e\n \u003cp\u003e82\u003c/p\u003e\n \u003cp\u003e\u0026nbsp;(78, 86)\u003c/p\u003e\n \u003cp\u003e\u0026nbsp; 307/373\u003c/p\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 68px;\"\u003e\n \u003cp\u003e63\u0026nbsp;\u003c/p\u003e\n \u003cp\u003e(54, 71)\u003c/p\u003e\n \u003cp\u003e\u0026nbsp; 78/124\u003c/p\u003e\n \u003cp\u003ep=0.420*\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 68px;\"\u003e\n \u003cp\u003e95\u0026nbsp;\u003c/p\u003e\n \u003cp\u003e(92, 98)\u0026nbsp;\u003c/p\u003e\n \u003cp\u003e\u0026nbsp;204/214\u003c/p\u003e\n \u003cp\u003ep=0.660*\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 60px;\"\u003e\n \u003cp\u003e89\u0026nbsp;\u003c/p\u003e\n \u003cp\u003e(80, 94)\u003c/p\u003e\n \u003cp\u003e\u0026nbsp; 78/88\u003c/p\u003e\n \u003cp\u003ep=0.416*\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 64px;\"\u003e\n \u003cp\u003e82\u003c/p\u003e\n \u003cp\u003e\u0026nbsp;(76, 86)\u003c/p\u003e\n \u003cp\u003e\u0026nbsp; 204/250\u003c/p\u003e\n \u003cp\u003ep=0.057*\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 84px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eFoam swab\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 68px;\"\u003e\n \u003cp\u003e65\u003c/p\u003e\n \u003cp\u003e\u0026nbsp;(58, 72)\u003c/p\u003e\n \u003cp\u003e\u0026nbsp; 123/188\u003c/p\u003e\n \u003cp\u003e\u003cstrong\u003ep=0.039\u003csup\u003e\u0026euro;\u003c/sup\u003e\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 68px;\"\u003e\n \u003cp\u003e92\u003c/p\u003e\n \u003cp\u003e\u0026nbsp;(89, 95)\u003c/p\u003e\n \u003cp\u003e\u0026nbsp; 321/349\u003c/p\u003e\n \u003cp\u003ep=0.100\u003csup\u003e\u0026euro;\u003c/sup\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 60px;\"\u003e\n \u003cp\u003e81\u003c/p\u003e\n \u003cp\u003e\u0026nbsp;(74, 87)\u003c/p\u003e\n \u003cp\u003e\u0026nbsp; 123/151\u003c/p\u003e\n \u003cp\u003ep=0.679\u003csup\u003e\u0026euro;\u003c/sup\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 89px;\"\u003e\n \u003cp\u003e83\u003c/p\u003e\n \u003cp\u003e\u0026nbsp;(79, 87)\u003c/p\u003e\n \u003cp\u003e\u0026nbsp; 321/386\u003c/p\u003e\n \u003cp\u003e\u003cstrong\u003ep=0.007\u003csup\u003e\u0026euro;\u003c/sup\u003e\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 68px;\"\u003e\n \u003cp\u003e57\u003c/p\u003e\n \u003cp\u003e\u0026nbsp;(47, 67)\u003c/p\u003e\n \u003cp\u003e\u0026nbsp; 58/101\u003c/p\u003e\n \u003cp\u003ep=0.385\u003csup\u003e\u0026euro;\u003c/sup\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 68px;\"\u003e\n \u003cp\u003e82\u003c/p\u003e\n \u003cp\u003e\u0026nbsp;(90, 97)\u003c/p\u003e\n \u003cp\u003e\u0026nbsp; 197/209\u003c/p\u003e\n \u003cp\u003ep=0.100\u003csup\u003e\u0026euro;\u003c/sup\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 58px;\"\u003e\n \u003cp\u003e83\u003c/p\u003e\n \u003cp\u003e\u0026nbsp;(72, 91)\u003c/p\u003e\n \u003cp\u003e\u0026nbsp; 58/70\u003c/p\u003e\n \u003cp\u003ep=0.617\u003csup\u003e\u0026euro;\u003c/sup\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 60px;\"\u003e\n \u003cp\u003e82\u003c/p\u003e\n \u003cp\u003e\u0026nbsp;(77, 87)\u0026nbsp;\u003c/p\u003e\n \u003cp\u003e\u0026nbsp;197/240\u003c/p\u003e\n \u003cp\u003ep=0.117\u003csup\u003e\u0026euro;\u003c/sup\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 68px;\"\u003e\n \u003cp\u003e74\u003c/p\u003e\n \u003cp\u003e\u0026nbsp;(63, 83)\u0026nbsp;\u003c/p\u003e\n \u003cp\u003e\u0026nbsp;62/84\u003c/p\u003e\n \u003cp\u003ep=0.164*\u003c/p\u003e\n \u003cp\u003e\u003cstrong\u003ep=0.013\u003csup\u003e\u0026euro;\u003c/sup\u003e\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 68px;\"\u003e\n \u003cp\u003e89\u0026nbsp;\u003c/p\u003e\n \u003cp\u003e(82, 94)\u003c/p\u003e\n \u003cp\u003e\u0026nbsp; 119/134\u003c/p\u003e\n \u003cp\u003ep=0.660*\u003c/p\u003e\n \u003cp\u003ep=0.100\u003csup\u003e\u0026euro;\u003c/sup\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 60px;\"\u003e\n \u003cp\u003e81\u003c/p\u003e\n \u003cp\u003e\u0026nbsp;(70, 89)\u003c/p\u003e\n \u003cp\u003e\u0026nbsp; 62/77\u003c/p\u003e\n \u003cp\u003ep=0.314*\u003c/p\u003e\n \u003cp\u003ep=0.906\u003csup\u003e\u0026euro;\u003c/sup\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 64px;\"\u003e\n \u003cp\u003e84\u003c/p\u003e\n \u003cp\u003e\u0026nbsp;(77, 90)\u003c/p\u003e\n \u003cp\u003e\u0026nbsp; 119/141\u003c/p\u003e\n \u003cp\u003e\u003cstrong\u003ep=0.005*\u003c/strong\u003e\u003c/p\u003e\n \u003cp\u003e\u003cstrong\u003ep=0.010\u003csup\u003e\u0026euro;\u003c/sup\u003e\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003c/tbody\u003e\n\u003c/table\u003e\n\u003cp\u003eWithin row p-values:\u0026nbsp;*HIV-negative vs. HIV-positive\u003c/p\u003e\n\u003cp\u003eWithin column p-values:\u0026nbsp;\u003csup\u003e\u0026euro;\u003c/sup\u003evs. flocked swab\u003c/p\u003e\n\u003cp\u003eAbbreviations: CI, confidence interval; eMRS, extended microbiological reference standard; HIV, human immunodeficiency virus; NPV, negative predictive value; OW, oral wash; PPV, positive predictive value; Ultra, Xpert MTB/RIF Ultra\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e\u003csup\u003e\u0026beta;\u003c/sup\u003e\u003c/strong\u003eTen people had unknown HIV status\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eTable 3:\u0026nbsp;\u003c/strong\u003eDiagnostic accuracy of Ultra on OW or tongue swabs compared to a double sputum culture as an eMRS for TB detection in Cohort B. Data are %, 95% CI, and n/N.\u003cstrong\u003e\u0026nbsp;\u003c/strong\u003e\u003c/p\u003e\n\u003ctable border=\"1\" cellspacing=\"0\" cellpadding=\"0\"\u003e\n \u003ctbody\u003e\n \u003ctr\u003e\n \u003ctd rowspan=\"2\" style=\"width: 147px;\"\u003e\n \u003cp\u003e\u003cstrong\u003e\u0026nbsp;\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"4\" style=\"width: 260px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eHead-to-head\u0026nbsp;\u003c/strong\u003e\u003c/p\u003e\n \u003cp\u003e\u003cstrong\u003e(n=122)\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"4\" style=\"width: 260px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eNon-head-to-head\u0026nbsp;\u003c/strong\u003e\u003c/p\u003e\n \u003cp\u003e\u003cstrong\u003e(n=156)\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 68px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eSensitivity\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 68px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eSpecificity\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 62px;\"\u003e\n \u003cp\u003e\u003cstrong\u003ePPV\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 62px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eNPV\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 68px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eSensitivity\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 68px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eSpecificity\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 62px;\"\u003e\n \u003cp\u003e\u003cstrong\u003ePPV\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 62px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eNPV\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 147px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eOral wash\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 68px;\"\u003e\n \u003cp\u003e56\u0026nbsp;\u003c/p\u003e\n \u003cp\u003e(21, 86)\u003c/p\u003e\n \u003cp\u003e5/9\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 68px;\"\u003e\n \u003cp\u003e89\u0026nbsp;\u003c/p\u003e\n \u003cp\u003e(82, 94)\u003c/p\u003e\n \u003cp\u003e101/113\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 62px;\"\u003e\n \u003cp\u003e29\u003c/p\u003e\n \u003cp\u003e\u0026nbsp;(10, 56)\u003c/p\u003e\n \u003cp\u003e5/17\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 62px;\"\u003e\n \u003cp\u003e96\u003c/p\u003e\n \u003cp\u003e\u0026nbsp;(91, 99)\u003c/p\u003e\n \u003cp\u003e101/105\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 68px;\"\u003e\n \u003cp\u003e71\u0026nbsp;\u003c/p\u003e\n \u003cp\u003e(42, 92)\u003c/p\u003e\n \u003cp\u003e10/14\u003c/p\u003e\n \u003cp\u003ep=0.435*\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 68px;\"\u003e\n \u003cp\u003e92\u0026nbsp;\u003c/p\u003e\n \u003cp\u003e(86, 96)\u003c/p\u003e\n \u003cp\u003e130/142\u003c/p\u003e\n \u003cp\u003ep=0.556*\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 62px;\"\u003e\n \u003cp\u003e45\u003c/p\u003e\n \u003cp\u003e\u0026nbsp;(24, 68)\u003c/p\u003e\n \u003cp\u003e10/22\u003c/p\u003e\n \u003cp\u003ep=0.307*\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 62px;\"\u003e\n \u003cp\u003e97\u003c/p\u003e\n \u003cp\u003e\u0026nbsp;(93, 99)\u003c/p\u003e\n \u003cp\u003e130/134\u003c/p\u003e\n \u003cp\u003ep=0.725*\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 147px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eSingle boiled tongue swab\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 68px;\"\u003e\n \u003cp\u003e67\u003c/p\u003e\n \u003cp\u003e\u0026nbsp;(30, 93)\u003c/p\u003e\n \u003cp\u003e6/9\u003c/p\u003e\n \u003cp\u003ep=0.065\u003csup\u003eΩ\u003c/sup\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 68px;\"\u003e\n \u003cp\u003e82\u0026nbsp;\u003c/p\u003e\n \u003cp\u003e(74, 89)\u003c/p\u003e\n \u003cp\u003e91/110\u003c/p\u003e\n \u003cp\u003ep=0.151\u003csup\u003eΩ\u003c/sup\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 62px;\"\u003e\n \u003cp\u003e23\u003c/p\u003e\n \u003cp\u003e\u0026nbsp;(9, 44)\u003c/p\u003e\n \u003cp\u003e6/26\u003c/p\u003e\n \u003cp\u003ep=0.695\u003csup\u003eΩ\u003c/sup\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 62px;\"\u003e\n \u003cp\u003e97\u003c/p\u003e\n \u003cp\u003e\u0026nbsp;(91, 99)\u003c/p\u003e\n \u003cp\u003e91/94\u003c/p\u003e\n \u003cp\u003ep=0.864\u003csup\u003eΩ\u003c/sup\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 68px;\"\u003e\n \u003cp\u003e64\u003c/p\u003e\n \u003cp\u003e\u0026nbsp;(35, 87)\u003c/p\u003e\n \u003cp\u003e9/14\u003c/p\u003e\n \u003cp\u003ep=0.907*\u003c/p\u003e\n \u003cp\u003ep=0.705\u003csup\u003eΩ\u003c/sup\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 68px;\"\u003e\n \u003cp\u003e81\u003c/p\u003e\n \u003cp\u003e\u0026nbsp;(74, 87)\u003c/p\u003e\n \u003cp\u003e114/140\u003c/p\u003e\n \u003cp\u003ep=0.79*\u003c/p\u003e\n \u003cp\u003e\u003cstrong\u003ep=0.018\u003c/strong\u003e\u003cstrong\u003e\u003csup\u003eΩ\u003c/sup\u003e\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 62px;\"\u003e\n \u003cp\u003e26\u0026nbsp;\u003c/p\u003e\n \u003cp\u003e(12, 43)\u003c/p\u003e\n \u003cp\u003e9/35\u003c/p\u003e\n \u003cp\u003ep=0.813*\u003c/p\u003e\n \u003cp\u003ep=0.143\u003csup\u003eΩ\u003c/sup\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 62px;\"\u003e\n \u003cp\u003e96\u003c/p\u003e\n \u003cp\u003e\u0026nbsp;(90, 99)\u003c/p\u003e\n \u003cp\u003e114/119\u003c/p\u003e\n \u003cp\u003ep=0.700*\u003c/p\u003e\n \u003cp\u003ep=0.609\u003csup\u003eΩ\u003c/sup\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 147px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eDouble tongue swab\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 68px;\"\u003e\n \u003cp\u003e22\u0026nbsp;\u003c/p\u003e\n \u003cp\u003e(3, 60)\u003c/p\u003e\n \u003cp\u003e2/9\u003c/p\u003e\n \u003cp\u003ep=0.160\u003csup\u003eΩ\u003c/sup\u003e\u003c/p\u003e\n \u003cp\u003ep=0.068\u003csup\u003e\u0026ne;\u003c/sup\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 68px;\"\u003e\n \u003cp\u003e96\u0026nbsp;\u003c/p\u003e\n \u003cp\u003e(91, 99)\u003c/p\u003e\n \u003cp\u003e108/112\u003c/p\u003e\n \u003cp\u003e\u003cstrong\u003ep=0.042\u003c/strong\u003e\u003cstrong\u003e\u003csup\u003eΩ\u003c/sup\u003e\u003c/strong\u003e\u003c/p\u003e\n \u003cp\u003e\u003cstrong\u003ep=0.009\u003c/strong\u003e\u003cstrong\u003e\u003csup\u003e\u0026ne;\u003c/sup\u003e\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 62px;\"\u003e\n \u003cp\u003e33\u003c/p\u003e\n \u003cp\u003e\u0026nbsp;(4, 78)\u003c/p\u003e\n \u003cp\u003e2/6\u003c/p\u003e\n \u003cp\u003ep=0.858\u003csup\u003eΩ\u003c/sup\u003e\u003c/p\u003e\n \u003cp\u003ep=0.639\u003csup\u003e\u0026ne;\u003c/sup\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 62px;\"\u003e\n \u003cp\u003e94\u003c/p\u003e\n \u003cp\u003e\u0026nbsp;(88, 98)\u003c/p\u003e\n \u003cp\u003e108/115\u003c/p\u003e\n \u003cp\u003ep=0.480\u003csup\u003eΩ\u003c/sup\u003e\u003c/p\u003e\n \u003cp\u003ep=0.395\u003csup\u003e\u0026ne;\u003c/sup\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 68px;\"\u003e\n \u003cp\u003e22\u0026nbsp;\u003c/p\u003e\n \u003cp\u003e(3, 60)\u003c/p\u003e\n \u003cp\u003e2/9\u003c/p\u003e\n \u003cp\u003e\u003cstrong\u003ep=0.022\u003c/strong\u003e\u003cstrong\u003e\u003csup\u003eΩ\u003c/sup\u003e\u003c/strong\u003e\u003c/p\u003e\n \u003cp\u003e\u003cstrong\u003ep=0.048\u003c/strong\u003e\u003cstrong\u003e\u003csup\u003e\u0026ne;\u003c/sup\u003e\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 68px;\"\u003e\n \u003cp\u003e96\u0026nbsp;\u003c/p\u003e\n \u003cp\u003e(91, 99)\u003c/p\u003e\n \u003cp\u003e108/112\u003c/p\u003e\n \u003cp\u003ep=0.117\u003csup\u003eΩ\u003c/sup\u003e\u003c/p\u003e\n \u003cp\u003e\u003cstrong\u003ep=0.0004\u003c/strong\u003e\u003cstrong\u003e\u003csup\u003e\u0026ne;\u003c/sup\u003e\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 62px;\"\u003e\n \u003cp\u003e33\u003c/p\u003e\n \u003cp\u003e\u0026nbsp;(4, 78)\u003c/p\u003e\n \u003cp\u003e2/6\u003c/p\u003e\n \u003cp\u003ep=0.595\u003csup\u003eΩ\u003c/sup\u003e\u003c/p\u003e\n \u003cp\u003ep=0.729\u003csup\u003e\u0026ne;\u003c/sup\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 62px;\"\u003e\n \u003cp\u003e94\u0026nbsp;\u003c/p\u003e\n \u003cp\u003e(88, 98)\u003c/p\u003e\n \u003cp\u003e108/115\u003c/p\u003e\n \u003cp\u003ep=0.247\u003csup\u003eΩ\u003c/sup\u003e\u003c/p\u003e\n \u003cp\u003ep=0.526\u003csup\u003e\u0026ne;\u003c/sup\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003c/tbody\u003e\n\u003c/table\u003e\n\u003cp\u003eWithin row p-values:\u0026nbsp;*Head-to-head vs. Non-head-to-head\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eWithin column p-values: \u003csup\u003eΩ\u003c/sup\u003evs. OW, \u003csup\u003e\u0026ne;\u003c/sup\u003evs. single boiled swab\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eAbbreviations: CI, confidence interval; eMRS, extended microbiological reference standard; NPV, negative predictive value; OW, oral wash;\u003cstrong\u003e\u0026nbsp;\u003c/strong\u003ePPV, positive predictive value; Ultra, Xpert MTB/RIF Ultra\u003cstrong\u003e\u0026nbsp;\u003c/strong\u003e\u003c/p\u003e"}],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":true,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":true,"hideJournal":true,"highlight":"","institution":"Stellenbosch University","isAcceptedByJournal":false,"isAuthorSuppliedPdf":false,"isDeskRejected":"","isHiddenFromSearch":false,"isInQc":false,"isInWorkflow":false,"isPdf":false,"isPdfUpToDate":true,"isWithdrawnOrRetracted":false,"journal":{"display":true,"email":"[email protected]","identity":"researchsquare","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":true,"externalIdentity":"","sideBox":"","snPcode":"","submissionUrl":"/submission","title":"Research Square","twitterHandle":"researchsquare","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"","reportingPortfolio":"","inReviewEnabled":false,"inReviewRevisionsEnabled":true},"keywords":"Xpert MTB/RIF Ultra, tuberculosis, oral washes, tongue swabs","lastPublishedDoi":"10.21203/rs.3.rs-6225530/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-6225530/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003cp\u003e\u003cu\u003eBackground\u003c/u\u003e: Oral samples show promise for tuberculosis (TB) diagnosis. Data from different samples and people with sputum scarce TB are limited.\u003c/p\u003e\n\u003cp\u003e\u003cu\u003eMethods:\u003c/u\u003e We assessed Xpert MTB/RIF Ultra (Ultra) in symptomatic people at clinics (Cohort A, n=891) or at antiretroviral therapy (ART)-initiation without syndromic preselection (Cohort B, n=258). In Cohort A, we collected oral washes (OWs) and, separately, tongue swabs (flocked, foam with heat). In Cohort B, we collected OWs, three flocked tongue swabs (comparing one with heat to two pooled swabs) and, separately, buccal swabs, periodontal brushes. We offered sputum induction and did different culture methods on a subset of Cohort B tongue swabs.\u003c/p\u003e\n\u003cp\u003e\u003cu\u003eResults:\u003c/u\u003e In Cohort A, Ultra on OWs, flocked tongue and foam swabs had sensitivities of 80% (95% confidence interval 56, 94), 59% (53, 65) and 65% (58, 72) and high specificities. In Cohort B, OWs and single heated swabs had 71% (42, 92) and 64% (35, 87) sensitivity, respectively. Pooled tongue swabs, buccal swabs and periodontal brushes had low sensitivities. MGIT960 had the highest sensitivity [64% (35, 87)] of culture methods. Oral sampling detected TB in sputum-scarce people [Cohort A: 25% (7/28) flocked and foam swab-positive; Cohort B: 18% (10/56) OW-, 23% (13/56) single flocked swab-positive]. In Cohort B, this would at least double the people with a positive Ultra result (sputum or oral) if induction were unavailable.\u003c/p\u003e\n\u003cp\u003e\u003cu\u003eConclusion:\u003c/u\u003e Ultra on OWs or foam tongue swabs has higher sensitivity than other oral-based approaches and detects sputum-scarce TB, resulted in more people diagnosed compared to Ultra on expectorated sputum.\u003c/p\u003e","manuscriptTitle":"Oral washes and tongue swabs for Xpert MTB/RIF Ultra-based tuberculosis diagnosis in people with and without the ability to make sputum","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2025-03-18 10:45:27","doi":"10.21203/rs.3.rs-6225530/v1","editorialEvents":[{"type":"communityComments","content":0}],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"researchsquare","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":true,"externalIdentity":"","sideBox":"","snPcode":"","submissionUrl":"/submission","title":"Research Square","twitterHandle":"researchsquare","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"","reportingPortfolio":"","inReviewEnabled":false,"inReviewRevisionsEnabled":true}}],"origin":"","ownerIdentity":"52fea32f-924c-4465-b6ad-bb9f01b8a9bc","owner":[],"postedDate":"March 18th, 2025","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"posted","subjectAreas":[{"id":45692032,"name":"Molecular Biology"}],"tags":[],"updatedAt":"2025-03-18T10:45:27+00:00","versionOfRecord":[],"versionCreatedAt":"2025-03-18 10:45:27","video":"","vorDoi":"","vorDoiUrl":"","workflowStages":[]},"version":"v1","identity":"rs-6225530","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-6225530","identity":"rs-6225530","version":["v1"]},"buildId":"8U1c8b4HqxoKbykW_rLl7","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}