Researchers Develop Rapid, Accurate Tuberculosis Test That Requires No Sputum Sample

On March 9, 2026, the World Health Organization issued its first-ever recommendation for tongue swab-based molecular testing for tuberculosis — a decision that, on paper, ends more than a century of dependence on sputum as the primary specimen for TB diagnosis [1]. The recommendation also endorsed near-point-of-care nucleic acid amplification tests (NPOC-NAATs), portable, battery-operated devices that deliver results in under an hour [1]. WHO Director-General Dr. Tedros Adhanom Ghebreyesus called the new tools "truly transformative," saying they bring "fast, accurate diagnosis closer to people, saving lives, curbing transmission and reducing costs" [2].

But TB has a long history of diagnostic breakthroughs that failed to reach the patients who needed them most. The question now is whether tongue swabs will follow a different path — or repeat the pattern.

The Sputum Problem

Sputum-based testing has been the backbone of TB diagnosis since Robert Koch identified Mycobacterium tuberculosis in 1882. Even the most advanced molecular test currently in widespread use — Cepheid's Xpert MTB/RIF, introduced in 2010 — requires a patient to cough up a thick mucus sample from deep in the lungs [3].

For many patients, that is impossible. Approximately 40% of adolescents and adults cannot produce a sputum sample even after careful coaching [4]. The figure is far worse for children under 10, who often require invasive procedures like gastric lavage — flushing the stomach with saline to recover swallowed sputum [5]. A meta-analysis of 15 studies covering 3,640 children found Xpert sensitivity of just 62% using expectorated sputum and 66% with gastric lavage, compared to 89% pooled sensitivity in adults [5].

HIV-positive patients face a parallel problem. Because HIV-associated TB tends to be paucibacillary — involving lower concentrations of bacteria — sputum samples from these patients often come back falsely negative [6]. This is not a marginal population: TB is the leading cause of death among people living with HIV worldwide [7].

The result is a massive diagnostic gap. In 2024, only 78% of estimated TB cases were formally diagnosed and notified, leaving roughly 2.4 million people with undiagnosed TB globally [8]. India, Indonesia, and the Philippines — the three highest-burden countries — together account for approximately 40% of that gap [8].

Source: WHO Global TB Report 2025

Data as of Oct 1, 2025CSV

What the Tongue Swab Tests Actually Do

The tongue swab approach is disarmingly simple: a healthcare worker runs a swab along the back of the patient's tongue, collecting oral bacteria including M. tuberculosis DNA shed from the lungs. The swab is then processed on existing molecular platforms [9].

Two commercial platforms have been validated for use with tongue swabs. Cepheid's Xpert MTB/RIF Ultra — already the most widely deployed molecular TB test — achieved 75.4% sensitivity and 100% specificity in a South African study of 321 participants [10]. Molbio Diagnostics' Truenat MTB Ultima, developed by its R&D subsidiary Bigtec Labs in Goa, India, showed 71.6% sensitivity and 96.9% specificity in the same study, with sensitivity rising to 79.1% when frozen samples were re-tested [10].

More impressive numbers have come from other study designs. A Rutgers University study using a sequence-specific magnetic capture (SSMaC) processing strategy on archived tongue swabs from 124 South Africans reported 90% sensitivity and 97% specificity [11]. A Ugandan study of 378 participants using a manual qPCR assay achieved 92.6% sensitivity and 99.1% specificity — performance the authors described as "rivaling sputum-based molecular TB diagnostics" [12]. A multi-country evaluation across four high-burden countries found tongue swab concordance rates of 95.1% compared to standard Xpert MTB/RIF sputum testing, with overall sensitivity of 88.6% and specificity of 98.3% [9].

Source: Multiple clinical studies

Data as of Mar 9, 2026CSV

The variation in sensitivity across studies reflects differences in processing methods, bacterial load in patients, and study design. Sensitivity tends to be higher in patients with higher bacterial loads — precisely the patients most likely to transmit TB to others [10].

The Burden Is Concentrated

TB kills over 3,300 people per day and produces more than 29,000 new cases daily [1]. The disease burden is heavily concentrated in a handful of countries.

Source: WHO Global TB Report 2025

Data as of Oct 1, 2025CSV

India alone accounts for an estimated 2.7 million cases — 25% of the global total. Indonesia contributes 1.08 million (10%), and the Philippines 0.73 million (6.8%) [8]. In the Philippines, incidence has risen 54% since 2000; in Indonesia, 35% [8]. These increases are driven by demographic growth, persistent transmission in crowded settings, and gaps in diagnostic coverage.

In India, the care cascade is especially troubled. Studies in Jharkhand and Gujarat found median total delay from symptom onset to treatment initiation of 55.3 days, with diagnostic delay accounting for 31 of those days [13]. In Mumbai, the mean total diagnostic pathway stretched to 65 days [14]. Nearly half of Indian TB patients first seek care from private practitioners, where diagnosis, treatment, and reporting are often substandard [13].

A faster, easier-to-collect test could shorten the diagnostic delay component. But treatment delay — the time between confirmed diagnosis and first dose of medication — is already relatively short in India, at roughly 2.5 to 7.4 days [13]. The larger bottlenecks are patient delay (time before seeking care), health system navigation, and supply-chain issues that affect drug availability at peripheral health posts.

Cost, Infrastructure, and the Access Question

The WHO announcement emphasized that NPOC-NAATs are available at "less than half the cost" of existing molecular diagnostics [1]. The current concessional price for Cepheid's Xpert MTB/RIF cartridge in high-burden developing countries is US$7.97, reduced from US$9.98 in September 2023 after sustained advocacy pressure [15]. Cepheid's drug-resistant TB cartridge (Xpert MTB/XDR) remains at US$14.90 [15].

These prices remain contested. A manufacturing cost analysis commissioned by Médecins Sans Frontières found that Cepheid could produce cartridges for as little as US$3 at scale, and MSF's "Time for $5" campaign argues the tests should be sold profitably at that price [16]. The campaign, supported by the Treatment Action Group and Partners in Health, has called Cepheid's pricing "inadequate" [17].

The instrument cost is a separate barrier. Molecular testing platforms range from US$17,530 to US$90,854, and a 2026 analysis from Boston University's School of Public Health found that at current pricing, fully decentralized TB testing is not cost-effective in most low-resource scenarios [18]. Molbio's Truenat platform, designed specifically for peripheral settings, has a lower instrument cost than GeneXpert, but exact figures vary by configuration and procurement channel.

Tongue swabs themselves add negligible cost — a standard swab is far cheaper than the sputum collection containers and processing supplies required for conventional testing. The real savings come from reduced need for trained sputum collection staff and the ability to test patients who would otherwise require referral to higher-level facilities.

Lessons from LAM: Why Past Breakthroughs Stalled

The TB diagnostics field has been burned before. The most prominent cautionary tale is the urine lipoarabinomannan (LAM) test, which promised point-of-care TB detection using a simple urine sample. In practice, LAM sensitivity ranged from 13% to 93% depending on the patient population and disease severity [19]. It worked best in severely immunocompromised HIV patients with CD4 counts below 100, but performance dropped sharply in HIV-negative patients and those with less advanced immunosuppression [19].

The next-generation FujiLAM (SILVAMP TB LAM) improved on the original, achieving 70.7% sensitivity in people with HIV and 53% in HIV-negative populations [19]. But scale-up was hampered by manufacturing inconsistencies, including variability between production lot numbers [19]. False positives linked to nontuberculous mycobacterial infections and cryptococcal disease further complicated field deployment [20].

The tongue swab approach has structural advantages over LAM. Rather than detecting a single biomarker with variable expression, it amplifies M. tuberculosis DNA using the same nucleic acid amplification technology that underpins GeneXpert and Truenat — platforms with established supply chains, training programs, and quality assurance systems in dozens of countries [9]. The specimen collection method changes; the underlying diagnostic technology does not. This means regulatory pathways, laboratory workflows, and quality control procedures can build on existing infrastructure rather than starting from scratch.

That said, WHO prequalification — the formal process by which WHO evaluates and approves diagnostic products for procurement by international agencies — is still in progress for tongue swab applications. In 2025, WHO launched a pilot parallel prequalification and policy assessment process for new TB in vitro diagnostics, with a submission deadline of August 30, 2025 [21]. The December 2024 prequalification of Xpert MTB/RIF Ultra was the first-ever WHO prequalification of a TB diagnostic test [22]. National regulatory approvals in high-burden countries will follow their own timelines.

The Research Pipeline Is Deep — and Under Threat

The tongue swab recommendation arrives during a period of intense research activity in TB diagnostics. According to OpenAlex data, publications on "tuberculosis diagnosis" peaked at 31,648 papers in 2023 and remained above 29,000 in 2025 [23]. The WHO reported approximately 100 diagnostic products in the development pipeline as of August 2025, alongside 29 drugs in clinical trials and 18 vaccine candidates in clinical development [24].

Source: OpenAlex

Data as of Jan 1, 2026CSV

But this research enterprise faces a funding crisis. Global TB research funding in 2023 totaled US$1.2 billion — just 24% of the US$5 billion annual target set by the 2023 UN high-level meeting on TB [24]. Diagnostics received 14% of that funding, behind drugs (35%) and vaccines (19%) [24]. Public-sector sources provided 62% of total TB research funding, with philanthropy contributing 24% and the private sector just 9% [24].

The proposed 40% cut to the US National Institutes of Health budget for 2026 — approximately US$18 billion — threatens to accelerate the shortfall [25]. Since early 2025, roughly 2,100 health research grants worth US$9.5 billion have been terminated, with an additional US$2.6 billion in research contracts cut [25]. In South Africa, 39 TB/HIV research sites are considered vulnerable, with 24 HIV trials and 20 TB trials at risk [25].

The Steelman Case for Skepticism

The most uncomfortable question about a faster, easier TB test is whether it might reduce urgency around harder problems. TB treatment regimens for drug-susceptible disease still require four to six months of daily medication. Drug-resistant TB treatment can last 9 to 20 months with severe side effects. Only three new TB drugs have reached approval in over 40 years: bedaquiline, delamanid, and pretomanid [24].

If a tongue swab test makes diagnosis feel "solved," governments and donors could plausibly redirect attention and funding away from the treatment pipeline, vaccine development, and healthcare worker capacity building. This concern is not hypothetical: after the introduction of GeneXpert in 2010, diagnostic access expanded while treatment completion rates in some settings remained stagnant or declined, partly because health systems could diagnose more patients than they could treat [26].

Dr. Tereza Kasaeva, Director of WHO's Department for HIV, TB, Hepatitis and STIs, addressed this obliquely in the March 2026 announcement, describing the new tools as "a major step forward" while emphasizing that diagnosis is only one component of the TB response [1]. The WHO's own data shows a return on TB investment of US$43 for every US$1 spent [2], suggesting that the economic case for comprehensive funding — diagnosis, treatment, and prevention — remains strong regardless of diagnostic advances.

There is also a practical counterargument: more diagnoses generate more political pressure, not less. Countries that find more TB cases face greater scrutiny from international bodies, domestic civil society, and their own public health targets. A test that reveals the true scale of the epidemic — including in populations previously invisible to the surveillance system — could force exactly the kind of health-system investment that skeptics worry will be deferred.

What Comes Next

The WHO recommendation is a necessary condition for widespread adoption, but it is not sufficient. The tongue swab must now clear several hurdles: national regulatory approvals in high-burden countries, procurement negotiations between manufacturers and international funders like the Global Fund, integration into national TB testing algorithms, and training of frontline health workers.

The intellectual property landscape is split between Cepheid (owned by Danaher Corporation) for Xpert Ultra and Molbio Diagnostics for Truenat. FIND, the Foundation for Innovative New Diagnostics, conducted the multi-centre diagnostic accuracy assessments that supported WHO endorsement [27]. Specific licensing terms for tongue swab applications in low- and middle-income countries have not been publicly disclosed.

The global TB diagnostics market is projected to grow from US$1.48 billion in 2025 to US$2.28 billion by 2033 [28]. Whether that growth translates into equitable access depends on pricing agreements that have historically been difficult to enforce. Sputum smear microscopy — a 19th-century technology — remains the primary diagnostic tool in many low- and middle-income countries because it is simple and cheap [28]. The tongue swab must compete not just with GeneXpert but with the economic reality that many health systems cannot afford molecular testing at all.

The most optimistic scenario: tongue swabs combined with portable NPOC-NAATs bring molecular TB testing to community health centers, schools, and HIV clinics that currently have no diagnostic capability beyond symptom screening. The most pessimistic: they become another tool available primarily in urban referral hospitals that already have GeneXpert machines, while the rural clinics serving the highest-burden populations continue to rely on clinical suspicion and empiric treatment.

The answer will be determined less by the science — which, for once, appears genuinely promising — than by the willingness of governments, manufacturers, and international funders to treat TB diagnostics as a public good rather than a commercial product.