health · family: it worked in the lab
resistant tb, detectedtoo far from care
Drug-resistant tuberculosis cannot be detected at the facilities where patients first seek care
Problem statement
Approximately 500,000 people develop drug-resistant tuberculosis each year, yet only one-third are diagnosed and enrolled on appropriate treatment. The gold-standard method — culture-based phenotypic drug susceptibility testing (DST) — can take up to two months for a full first-line and second-line resistance profile, requires infrastructure and technical skills rarely available outside reference laboratories, and is unavailable at the peripheral health facilities where most patients first present. GeneXpert MTB/RIF detects only rifampicin resistance, providing no guidance for selecting alternative regimens when resistance to isoniazid, fluoroquinolones, or newer drugs like bedaquiline is present. The result: most patients with drug-resistant TB are either never diagnosed or are treated with ineffective regimens for weeks before resistance is identified.
Why this matters
Delayed or incorrect treatment drives ongoing transmission of resistant strains, creating a positive feedback loop. The WHO TPP identifies four priority drugs for peripheral DST — rifampicin, isoniazid, fluoroquinolones, and bedaquiline — because knowing resistance status for these four enables selection of the appropriate treatment regimen. No existing test covers all four at the peripheral level.
What’s been tried and why it hasn’t worked
GeneXpert MTB/RIF (2010) was a breakthrough for rifampicin resistance detection but tests only one drug. GeneXpert MTB/XDR expands coverage to isoniazid, fluoroquinolones, second-line injectables, and ethionamide, but does not test for bedaquiline. Whole genome sequencing offers comprehensive resistance profiling but has incomplete knowledge of resistance mutations for newer drugs. Line probe assays (Hain GenoType) cover first- and second-line drugs but require specialized infrastructure and skills. Critically, none of these approaches delivers results fast enough for same-day treatment decisions: the WHO TPP optimal target is <30 minutes for detection plus DST.
What would unlock progress
Two approaches show promise: (1) microfluidic chip-based phenotypic DST, which reads resistance from bacterial growth in the presence of each drug (the cited sources do not report results for it), and (2) targeted next-generation sequencing panels that could cover all priority resistance mutations on a portable platform. The WHO TPP sets the bar: instrument cost <$5,000 (optimal), time to result <30 minutes (optimal) or <6 hours (minimum), deployable at peripheral health centers. Bridging this gap requires simultaneous advances in microfluidics, sample processing, and cost reduction.
Entry points for student teams
A team could prototype a microfluidic DST chip optimized for the four WHO priority drugs, using fluorescence-based readout of mycobacterial growth in drug-containing vs. drug-free chambers, with the key engineering challenge being reduction of incubation time to <6 hours while maintaining sensitivity — but developed on a surrogate organism rather than M. tuberculosis: Mycobacterium smegmatis mc(2)155 (ATCC 700084, biosafety level 1) for the fluidics, growth readout and time-to-detection work, or M. tuberculosis variant bovis BCG (ATCC 35734, biosafety level 2) where a slow-growing complex member matters. Confirmation on clinical M. tuberculosis isolates is the one step a student team cannot do for itself — it needs BSL-3 containment and banked patient isolates, held by national TB reference laboratories, the WHO supranational reference laboratory network, and the handful of universities with mycobacteriology BSL-3 cores — so the semester deliverable is a characterized chip plus the validation protocol written for such a lab. A second door needs no laboratory at all: design the targeted sequencing panel for rifampicin, isoniazid, fluoroquinolone and bedaquiline resistance and validate it in silico against genuinely public data — WHO's Catalogue of mutations in Mycobacterium tuberculosis complex and their association with drug resistance, 2nd ed. (2023) publishes graded mutation–resistance associations as a free downloadable spreadsheet, and the CRyPTIC consortium's reuse table (`ftp.ebi.ac.uk/pub/databases/cryptic/release_june2022/reuse/`) pairs isolate genotypes with measured MICs across 13 drugs, together enough to compute the panel's sensitivity, name its coverage gaps for bedaquiline, and benchmark cost and turnaround against GeneXpert XDR. Relevant disciplines: microfluidics, biomedical engineering, molecular biology, bioinformatics.
Genes
Structural cousins: stuck the same way, in other fields
Sources
WHO, "Target product profile for next-generation drug-susceptibility testing at peripheral centres," 2021, ISBN 9789240032361. MacLean EL-H et al., "Updating the WHO target product profile for next-generation Mycobacterium tuberculosis drug susceptibility testing at peripheral centres," PLOS Global Public Health, 2023. go to source 1 ↗ go to source 2 ↗
Working notes
Notes made while writing and checking this brief, shown as written. They use working shorthand.
- Companion problem to health-tb-poc-diagnostic-sputum-barrier: that brief addresses TB detection; this one addresses resistance profiling. Both share the constraint of peripheral deployment.
- The microfluidic chip approach (phenotypic DST) is a rare example of a technology that could leapfrog molecular approaches by directly measuring growth inhibition, avoiding the genotype-phenotype discordance problem.
- Bedaquiline resistance is particularly urgent: it is the backbone of new MDR-TB regimens, but resistance is already emerging and no rapid molecular test reliably detects it.
- The `failure:tech-limitation-now-resolved` tag reflects that the inability to detect any drug resistance at point of care was resolved by GeneXpert MTB/RIF (2010, rifampicin) and MTB/XDR (isoniazid + fluoroquinolones + second-line injectables). The broader problem persists: comprehensive multi-drug DST at peripheral level in <30 minutes remains unsolved.
Reconciliation 2026-08-21: Entry-point repair under the 2026-08-21 realism rules. The triage flag is CONFIRMED: both original doors silently required work on M. tuberculosis itself — a phenotypic growth chip run on the pathogen needs BSL-3 containment, and the "portable, low-cost targeted sequencing panel" as written needed clinical isolates to benchmark against GeneXpert XDR — so a student team had no way in. The chip door now specifies surrogate organisms and carries an explicit access line for the step that genuinely needs containment (BSL-3 plus banked isolates, held by national TB reference laboratories, the WHO supranational reference laboratory network, and universities with mycobacteriology BSL-3 cores), with the semester deliverable being the characterized chip plus a validation protocol for such a lab. The sequencing-panel door was converted to a facility-free in-silico validation. Resources verified by fetch on 2026-08-21: WHO, Catalogue of mutations in Mycobacterium tuberculosis complex and their association with drug resistance, 2nd ed., 2023, ISBN 9789240082410 — free PDF plus the full catalogue as a downloadable Excel file, CC BY-NC-SA 3.0 IGO (https://www.who.int/publications/i/item/9789240082410); the CRyPTIC consortium reuse tables, open on the EBI FTP with no application or login (CRyPTIC_reuse_table_20240917.csv and earlier releases, plus README/CHANGELOG, at http://ftp.ebi.ac.uk/pub/databases/cryptic/release_june2022/reuse/). Biosafety characterizations were checked against ATCC's own catalogue records rather than assumed: M. smegmatis mc(2)155 is ATCC 700084 and is listed at BSL-1, not BSL-2 as the triage row had it (https://www.atcc.org/products/700084), while M. tuberculosis variant bovis BCG strain TMC 1011 is ATCC 35734 at BSL-2 (https://www.atcc.org/products/35734) — the brief states each level as ATCC assigns it. No other section was touched.
Reconciliation 2026-09-29: The whole write-up was read against both cited sources, and both were read in full on 2026-09-29. The WHO target product profile (52 pages) was downloaded as a PDF from https://iris.who.int/server/api/core/bitstreams/88503945-2a26-4f49-969b-7fdf09405193/content, which is the download link on the cited WHO page. The journal article was read at the cited PLOS address. (1) The Source line gave the article's authors as "Kohli M et al." The article's own citation line begins "MacLean EL-H, Miotto P, González Angulo L", and Kohli is a later author. The Source line now says "MacLean EL-H et al." (2) The text said culture-based testing "takes 2-6 weeks" and "requires BSL-3 laboratory infrastructure." Neither source gives those figures. The article says the method can take up to two months for a full first-line and second-line profile, and that it is rarely available outside reference laboratories because of infrastructure and technical skill requirements. The sentence now says that. (3) The text said drug-resistant TB kills approximately 150,000 people per year, and that MDR-TB treatment takes 9-18 months and costs 10-100x more than treatment of drug-susceptible TB. Neither source contains these figures, so both sentences were removed. (4) The text gave GeneXpert MTB/XDR a date of 2021 and "~93% sensitivity and 98% specificity for fluoroquinolone resistance," and said it "runs on the same infrastructure-dependent platform." Neither source gives the date, the accuracy figures, or the statement about the platform. Both sources say the assay tests for resistance to isoniazid, fluoroquinolones, second-line injectables and ethionamide, and the sentence now says that and notes that bedaquiline is not on the list. The date was also removed from the Source Notes bullet about the tag. (5) The text said whole genome sequencing "costs >$100/test" and "requires specialized equipment." Neither source says so, and both phrases were removed. (6) The text said line probe assays "require BSL-3 facilities and trained molecular biologists." The WHO document says the first such tests "required specialized infrastructure and skills," and the sentence now uses those words. (7) The text said a microfluidic chip "has demonstrated growth-based resistance detection in 12 hours for seven drugs including bedaquiline and levofloxacin." Neither source mentions microfluidic chips or this result. The sentence now describes the approach and says the cited sources do not report results for it. The 12-hour figure was also removed from the entry-points section and from the Source Notes bullet. The suggestion itself is unchanged. Confirmed against the sources and left as written: about 500,000 drug-resistant cases and one third diagnosed and enrolled on treatment (article abstract, for 2020); the four priority drugs rifampicin, isoniazid, fluoroquinolones and bedaquiline, chosen to enable selection of the treatment regimen (WHO document); no WHO-endorsed molecular test for bedaquiline yet (article); time to result under 30 minutes as the optimal target and under 6 hours as the minimal one; instrument cost under US$ 5000 as the optimal target; delays in diagnosis and treatment fuel onward transmission (WHO document); incomplete knowledge of resistance mutations for newer drugs; targeted sequencing on small instruments; and the WHO title, year and ISBN. Left as written but not confirmed: the 2010 date for GeneXpert MTB/RIF, which neither source states directly (the WHO reference list cites a 2010 paper on the test); the phrase "positive feedback loop," which is this write-up's wording for what the sources say about transmission and amplified resistance; and the statement that patients are treated with ineffective regimens "for weeks." The entry-point resources were checked in the note above and were not checked again. Genes untouched.