The scale of the problem is considerable. Globally, an estimated 13 percent of deaths attributed to antimicrobial resistance are due to drug-resistant tuberculosis, much of which remains undetected. In 2024, roughly 390,000 people were estimated to have developed rifampicin-resistant or multidrug-resistant TB, defined as resistance to both rifampicin and isoniazid, the two most critical first-line anti-TB drugs. Missed or erroneous detection of resistance creates gaps in surveillance and can lead to mismanagement of treatment, perpetuating transmission and driving the epidemic.

For decades, culture-based phenotypic drug susceptibility testing has been considered the gold standard for detecting resistance, but the method has known limitations, including inaccuracy in detecting particular cases of resistance to rifampicin. Genotypic tests, such as the Xpert MTB/RIF assay and line probe assays, have accelerated the turnaround time for confirming drug-resistant TB, but they interrogate only narrow regions within a limited number of resistance-associated genes. Sequencing-based detection offers a more comprehensive approach and is included in global guidance for programmatic implementation, yet practical barriers related to cost, laboratory infrastructure, and technical expertise have hindered adoption in many high-burden countries.

Nigeria, a high HIV/TB and multidrug-resistant TB burden country, prioritized addressing DR-TB in its National Strategic Plan for 2021 to 2025. Under the leadership of the National Tuberculosis and Leprosy Control Programme, and with support from the U.S. Centers for Disease Control and Prevention, the country established its first sequencing model for DR-TB surveillance in 2022. The approach deliberately leveraged sequencing laboratories previously established for HIV drug-resistance monitoring, integrating services for these pathogens for the first time and demonstrating how limited resources can be stretched through a multi-disease platform.

The technical workflow began at zonal and national TB reference laboratories, where clinical sputum specimens from health facilities across the country are routinely subjected to genotypic tests. Drug-resistant isolates are then referred for phenotypic testing using the automated BACTEC MGIT 960 system. For the sequencing pilot, 49 Mycobacterium tuberculosis isolates previously identified as resistant to rifampicin and/or isoniazid were selected proportionately across all six geopolitical zones. Mycobacterial cultures were inactivated using heat at 95 degrees Celsius for 20 minutes and ultrasonication for 15 minutes, and genomic DNA was extracted on site before being shipped through a courier-based referral pathway to sequencing laboratories, where quality and quantity were assessed using spectrophotometry and fluorometry.

Rather than sequencing whole genomes, the team designed a targeted Sanger sequencing algorithm covering eight genomic regions associated with resistance to six anti-TB drugs, a strategy intended to maximize programmatic value while keeping costs manageable. The rpoB gene, linked to rifampicin resistance, was sequenced for all isolates to confirm results or identify resistance missed by other tests. Rv0678 and atpE, associated with resistance to the newer drugs bedaquiline and clofazimine, were also sequenced for all isolates. Genes including gyrA, gyrB, katG, the inhA promoter, and rrs were sequenced conditionally, only for isolates determined to be susceptible to the corresponding drug classes, in order to detect mutations not targeted by line probe assays. Identified mutations were cross-referenced with the World Health Organization’s global catalogue of resistance-associated mutations.

Implementation proceeded in four phases over two years: planning, validation of the new method, rollout, and data analysis for programmatic use. Laboratory staff received training, standard operating procedures were developed, and sequencing-specific quality control measures and performance indicators were established. Validation was successful at all participating laboratories, which exceeded established acceptance criteria. Of the 49 isolates selected, genomic DNA suitable for sequencing was extracted from 96 percent, adherence to the sequencing algorithm reached 100 percent, and high-quality results were obtained for 623 of 642 sequencing reactions, or 97 percent.

The results demonstrated the added value of sequencing in striking terms. While sequencing largely confirmed resistance reported by routine tests, it identified additional rifampicin resistance in 5 of 47 isolates, or 11 percent, and additional isoniazid resistance in 1 isolate, or 2 percent, that had been missed by phenotypic or genotypic testing. Discordance between the two routine methods was also notable, observed for rifampicin in 9 percent of isolates, isoniazid in 15 percent, fluoroquinolones in 9 percent, and second-line injectables in 6 percent. In several cases, sequencing settled these disagreements, confirming rpoB mutations that one method had detected and the other had not.

Particularly significant were nine borderline mutations in rpoB associated with low-level rifampicin resistance, identified in isolates from all geopolitical zones. Phenotypic testing can fail to detect these clinically relevant variants, which may fuel transmission of resistance if left undiagnosed. The study found that phenotypic testing missed rifampicin resistance in two isolates with confirmed borderline mutations, L452P and L430P, findings consistent with comparative studies from other high-burden settings. The analysis also revealed the limits of automated bioinformatics software: one potential instance of rifampicin heteroresistance, indicated by a double chromatogram peak, and a unique insertion mutation in rpoB were missed by automated tools, which are often restricted to a database of defined mutations.

Not every result favored sequencing. No additional resistance was detected in genes associated with fluoroquinolone or second-line injectable resistance among isolates deemed susceptible by routine tests, and no mutations were found in Rv0678 or atpE even among the small number of isolates phenotypically resistant to bedaquiline or clofazimine, underscoring that phenotypic testing remains essential for newer drugs whose resistance mechanisms are not yet fully understood. The authors also acknowledge limitations, including the small sample size and algorithm constraints that restrict generalizability, and the inability to repeat phenotypic testing due to resource restrictions.

Following the analysis, the National Tuberculosis and Leprosy Control Programme convened a stakeholder meeting to review lessons learned and chart the path forward. Leveraging existing laboratory infrastructure and workforce expertise, along with the targeted algorithm design, supported cost-effective implementation. Building on the pilot’s success, the program is now transitioning from Sanger sequencing at HIV laboratories to targeted next-generation sequencing at the Nigeria Centre for Disease Control and Prevention’s multi-disease testing laboratory. Next-generation sequencing offers higher throughput and improved sensitivity for detecting heteroresistance at lower frequencies, and commercially available end-to-end solutions with automated bioinformatic pipelines may expand uptake. The work demonstrates feasibility and provides a practical model for operationalizing DR-TB sequencing in a national program setting, offering an example for other high-burden countries seeking to strengthen genomic surveillance, improve drug selection and treatment for TB patients, and guide testing policies as new treatment regimens are scaled up.

Subject of Research: Implementation of DNA sequencing-based surveillance for drug-resistant tuberculosis in Nigeria

Article Title: Implementation of sequencing-based surveillance for drug-resistant tuberculosis in Nigeria

Article References: McLaughlin, H. P., Audu, I., Tola, M., Onwuamah, C. K., Ogwu, J., Atoe, K., Kunle-Ope, C. N., Ahmadu, I., Gowon, J., Meshak, D. B., Iwakun, M., Ude, C. E., Abiona, M., Momoh, A. E., Ore, T., Ogbu, C. A., Abiodun, T. A., Dawson, J., Hall-Eidson, P., … Chukwuma, A. (2026). Implementation of sequencing-based surveillance for drug-resistant tuberculosis in Nigeria. Frontiers in Tuberculosis, 4, Article 1855903. https://doi.org/10.3389/ftubr.2026.1855903

Image Credits: AI Generated

DOI: 10.3389/ftubr.2026.1855903

Keywords: tuberculosis, drug resistance, genomic surveillance, Sanger sequencing, Nigeria, Mycobacterium tuberculosis, rifampicin, isoniazid, drug susceptibility testing, next-generation sequencing, public health, antimicrobial resistance

The stakes of this question are far from academic. Fluoroquinolones, including drugs such as levofloxacin and moxifloxacin, form the backbone of the shorter, all-oral treatment regimens that the World Health Organization now endorses for patients whose tuberculosis does not respond to rifampicin, a key first-line antibiotic. When a physician knows that a patient’s tuberculosis strain is susceptible to fluoroquinolones, a shorter regimen becomes a realistic option, sparing the patient months of toxic therapy. When the strain is resistant, that regimen would fail, and an alternative must be assembled. The problem is that in many of the places where drug-resistant tuberculosis burdens are heaviest, laboratory capacity for rapid drug susceptibility testing—the gold standard for detecting resistance—is limited or absent, leaving clinicians to make consequential decisions with incomplete information.

Prediction models have emerged as one proposed bridge across this diagnostic gap. The idea is elegant in principle: by feeding routinely collected patient information—age, sex, treatment history, comorbidities, social risk factors—into a statistical algorithm, a model could estimate the probability that a patient’s tuberculosis is already resistant to fluoroquinolones, even without laboratory confirmation. Clinicians could then weigh that probability when selecting a regimen, much as they weigh other forms of clinical judgment. But most prediction models in this field have been developed and evaluated within a single country, using data from one health system and one circulating population of tuberculosis strains. Whether such models generalize to different epidemiological settings has remained an open and largely untested question.

To answer it, the research team—Tianfang Shao, Mariana R. Neves, Molly Franke, Carole Mitnick, Jennifer Furin, Ted Cohen, and Reza Yaesoubi—turned to the TB Portals, an open-access data-sharing platform curated by the National Institute of Allergy and Infectious Diseases. From this resource, they assembled records for 5,175 patients with rifampicin-resistant tuberculosis who had documented fluoroquinolone drug susceptibility testing results, drawn from eight countries: Azerbaijan, Belarus, Georgia, Kazakhstan, Kyrgyzstan, Moldova, Romania, and Ukraine, with data collected between 2012 and 2024. The burden in this cohort was substantial: 1,772 patients, or 34.2 percent, carried tuberculosis strains that were already resistant to fluoroquinolones. That high prevalence reflects both the severity of the drug resistance epidemic in Eastern Europe and Central Asia and the clinical urgency of identifying resistant cases before committing patients to a regimen that depends on these drugs.

Methodologically, the study was designed to stress-test generalizability in a way that few previous analyses have attempted. The researchers built prediction models using three distinct algorithms spanning a range of complexity: logistic regression, a transparent and widely used statistical technique; neural networks, which can capture intricate nonlinear relationships among predictors; and XGBoost, a gradient-boosted decision tree method that has become a favorite in clinical prediction research for its flexibility and performance. Each algorithm was then evaluated under three different training and validation strategies. The first strategy pooled data from all eight countries to train a single multi-country model. The second trained and tested models within individual countries using internal validation. The third—and most revealing—trained models on data from a subset of countries and then externally validated them on countries deliberately held out of training, simulating exactly what happens when a model built elsewhere is imported into a new setting.

Model performance was quantified using two standard metrics. The area under the receiver operating characteristic curve, or AUROC, measures how well a model separates resistant from susceptible cases across all possible decision thresholds, with 0.5 representing chance performance and 1.0 representing perfect discrimination. The area under the precision-recall curve, or AUPRC, is often considered more informative when the outcome of interest is relatively uncommon, because it focuses on how precisely the model identifies true positive cases. The pooled, multi-country models achieved what the authors describe as moderate discrimination after optimism correction, with AUROC values ranging from 0.70 to 0.72 and AUPRC values from 0.57 to 0.59 across the three algorithms. In practical terms, such models carry real information, but they would misclassify a substantial fraction of patients if used alone to route individuals toward or away from fluoroquinolone-based regimens.

Within-country models performed somewhat better, with AUROC and AUPRC values reaching 0.8 for some countries, suggesting that locally trained models can extract more signal from their own patient populations than a globally pooled model can. The cross-country external validation, however, exposed the fragility of geographic transfer. When a model trained on external data was applied to a held-out country, the loss in performance ranged from negligible to more than 0.1 in AUROC or AUPRC, depending on the country and the algorithm involved. A drop of that magnitude is clinically consequential: it can mean the difference between a model that usefully flags resistant infections and one that generates enough false reassurance to steer patients toward regimens destined to fail. The variability across settings indicates that no single number can capture how well a model will travel, and that each new deployment demands its own rigorous evaluation.

Just as striking as the performance differences were the differences in which predictors mattered. A limited set of variables—chiefly those related to the case definition and to a patient’s treatment history—emerged as the most consistently informative predictors across countries and algorithms. This makes biological and epidemiological sense: patients who have been treated for tuberculosis before, particularly with fluoroquinolones or other second-line drugs, face a higher likelihood of harboring strains that have already evolved resistance. By contrast, demographic characteristics, comorbidities, social risk factors, education, and employment variables contributed far more inconsistently, with their predictive weight swinging widely across countries and modeling approaches. This instability suggests that the social and structural determinants of drug resistance are woven into local contexts in ways that a model trained elsewhere cannot simply assume.

The authors are careful to note an important limitation of their analysis. The incidence of rifampicin-resistant tuberculosis and of fluoroquinolone resistance was relatively stable across the years represented in their dataset, which means their findings may not extend to settings where the dynamics of multidrug-resistant tuberculosis are changing markedly—for example, where new regimens are rapidly reshaping patterns of resistance. Prediction models are, at their core, pattern-recognition devices calibrated to the populations and time periods that produced their training data, and shifts in transmission, treatment practices, or drug exposure can quietly invalidate the relationships a model has learned. Any real-world deployment would therefore need ongoing monitoring and periodic recalibration, not just a one-time validation.

The study’s conclusions carry a clear message for the field of tuberculosis care and for clinical prediction modeling more broadly. Predicting fluoroquinolone resistance from demographic and clinical characteristics is feasible to a moderate degree, and locally developed models can offer meaningful support to clinicians working without rapid laboratory testing. But models developed for one country, or even for several countries pooled together, cannot be assumed to perform elsewhere without rigorous external validation in the destination setting. As all-oral, shorter regimens spread across the globe and diagnostic capacity lags behind, the researchers argue that the path forward lies in locally informed prediction models—tools built and validated on the populations they are meant to serve—rather than in the export of algorithms whose accuracy quietly dissolves at national borders.

Subject of Research: Cross-country validation of prediction models for fluoroquinolone resistance in rifampicin-resistant tuberculosis

Article Title: Predicting resistance to fluoroquinolones among patients with rifampicin-resistant tuberculosis: A cross-country validation study

Article References: Shao, T., Neves, M. R., Franke, M., Mitnick, C., Furin, J., Cohen, T., & Yaesoubi, R. (2026). Predicting resistance to fluoroquinolones among patients with rifampicin-resistant tuberculosis: A cross-country validation study. PLOS Medicine, 23(9), e1004965. https://doi.org/10.1371/journal.pmed.1004965

Image Credits: AI Generated

DOI: 10.1371/journal.pmed.1004965

Keywords: tuberculosis, fluoroquinolone resistance, rifampicin-resistant tuberculosis, prediction models, drug susceptibility testing, machine learning, XGBoost, external validation, TB Portals, WHO treatment regimens, multidrug-resistant tuberculosis, clinical decision-making

The research focused on a population of domestic dogs from an endemic rural area, sampling between July and October 2025. Twenty naturally Uncinaria-infected dogs and two puppies were enrolled, alongside three uninfected dogs that served as comparators. Fecal and blood samples were collected repeatedly, with egg shedding quantified on days minus 14, minus 7, and day 0 using the Mini-FLOTAC technique, a flotation-based counting method prized for its accuracy in quantifying parasite eggs per gram of feces. At day 0, the dogs were treated with one of three common anthelmintics: fenbendazole, moxidectin, or milbemycin oxime. Efficacy was then measured with the fecal egg count reduction test, or FECRT, the gold-standard field method that compares egg counts before and after treatment, in this case at day 14 post-treatment.

One of the most striking findings concerns the sheer volatility of egg shedding before any drug was administered. Egg counts ranged from zero to 3,105 eggs per gram across the pre-treatment sampling window, with sharp swings between consecutive sampling days. The mean egg count rose from 122 eggs per gram at day minus 14 to 647.2 at day minus 7, only to fall back to 162.2 at day 0. Such dramatic fluctuation within a matter of days carries a sober practical message: a single fecal sample can badly misrepresent the true infection status of a dog. A negative or low count on one day might mask a substantial infection that would be revealed a week later, which is precisely why the authors argue for multi-sample fecal diagnostics in both clinical practice and surveillance programs.

When it came to treatment, the results split cleanly. Fenbendazole, a member of the benzimidazole class that disrupts the parasite’s microtubule machinery by binding to β-tubulin, achieved a 100 percent fecal egg count reduction. Moxidectin, a milbemycin-class macrocyclic lactone that interferes with glutamate-gated chloride channels in the worm’s nervous system, came close behind at 99.3 percent efficacy. Milbemycin oxime, however, managed only 76.8 percent reduction, a figure that falls short of the thresholds typically used to declare acceptable efficacy and that raises questions about how this drug performs against U. stenocephala under field conditions. Importantly, the researchers did not stop at counting eggs; they probed the molecular and phenotypic basis of drug susceptibility.

To assess benzimidazole resistance, the team performed an in vitro egg hatch assay, exposing hookworm eggs to increasing concentrations of a benzimidazole and measuring the concentration that inhibits hatching in half of the eggs, the IC50. Two IC50 values were recorded, 0.03 with a 95 percent confidence interval of 0.03 to 0.04, and 0.09 with a confidence interval of 0.08 to 0.10, both consistent with phenotypic susceptibility to the drug class. In other words, despite the modest performance of milbemycin oxime in the live animals, the benzimidazole pathway itself showed no functional signs of resistance in this population.

Molecular screening reinforced that conclusion. Using amplicon deep sequencing of the internal transcribed spacer 2 region of the ribosomal RNA gene, the researchers confirmed that U. stenocephala was the sole hookworm species present in the samples, ruling out hidden co-infections with other hookworms such as Ancylostoma species that could confound efficacy calculations. They then sequenced the β-tubulin isotype-1 gene, the canonical molecular target whose mutations are the best-validated markers of benzimidazole resistance in parasitic nematodes. Across codons 134, 167, 198, and 200, the positions most frequently implicated in resistance, no canonical resistance-associated single nucleotide polymorphisms were detected, either before or after treatment. The genetic evidence and the phenotypic egg hatch data thus tell a consistent story: this southern Italian population of U. stenocephala remains fully susceptible to benzimidazoles.

The clinical pathology component of the study adds a reassuring but nuanced dimension. Hematological and biochemical values in the infected dogs were mostly within reference limits, a profile consistent with low-intensity parasitism rather than clinically significant disease. The most frequent abnormalities were mild eosinophilia, the elevated eosinophil count typical of the immune response to helminths; slight hypoalbuminemia, a modest drop in the blood protein albumin that hookworms can induce through intestinal blood loss and protein leakage; and increased β-globulin fractions, reflecting the antibody-driven immune activation that chronic parasitic exposure produces. None of these findings amounted to the severe anemia or profound protein loss that can accompany heavy hookworm burdens, suggesting that in these naturally exposed dogs, U. stenocephala was more of a subclinical passenger than a dangerous pathogen.

That subclinical profile is itself part of the reason the parasite has flown under the radar. U. stenocephala, sometimes called the northern hookworm, is better adapted to cooler climates than its tropical and subtropical relatives, and its infections tend to produce lower egg outputs and milder clinical signs. Yet the parasite is being reported with increasing frequency in Europe, and without baseline data on its shedding dynamics, drug susceptibility, and clinical impact, veterinarians have had little ground to stand on when interpreting diagnostic results or choosing treatments. This study supplies exactly that baseline, and its message about the unreliability of single-sample diagnosis is likely to resonate well beyond the study region.

The work also carries broader significance for the global problem of anthelmintic resistance. In canine hookworms, particularly Ancylostoma caninum in North America, widespread resistance to multiple drug classes has become a serious clinical challenge, making the search for populations where susceptibility persists all the more valuable. By documenting, with molecular and phenotypic rigor, a European U. stenocephala population that still responds fully to fenbendazole and moxidectin, the researchers have established a reference point against which future changes can be measured. Continued surveillance of anthelmintic efficacy, they emphasize, is essential, because resistance mutations can spread rapidly once they arise, and the modest efficacy of milbemycin oxime observed here warrants closer attention in future studies.

For dog owners and veterinarians, the practical takeaways are concrete. Repeated fecal examinations over consecutive weeks give a far more accurate picture of hookworm infection than a single test, given the day-to-day variability in egg shedding documented here. Fenbendazole and moxidectin remain dependable choices against U. stenocephala in this region, while milbemycin oxime’s performance suggests it should not be assumed fully effective without verification. And as this underestimated hookworm continues its apparent expansion across Europe, the integrated approach demonstrated in Basilicata, pairing egg counts with clinical chemistry, deep amplicon sequencing, and egg hatch assays, offers a template for monitoring not just this species, but the wider and shifting landscape of canine parasitism in a changing world.

Subject of Research: Anthelmintic susceptibility and clinical impact of the canine hookworm Uncinaria stenocephala in naturally infected dogs

Article Title: Clinical pathology, parasitological evaluation, and anthelmintic resistance assessment in dogs naturally infected with Uncinaria stenocephala

Article References: Fagundes-Moreira, R., Bezerra-Santos, M. A., Šlapeta, J., & Otranto, D. (2026). Clinical pathology, parasitological evaluation, and anthelmintic resistance assessment in dogs naturally infected with Uncinaria stenocephala. Parasites & Vectors. https://doi.org/10.1186/s13071-026-07688-4

Image Credits: AI Generated

DOI: 10.1186/s13071-026-07688-4

Keywords: Uncinaria stenocephala, hookworms, dogs, anthelmintic resistance, benzimidazole, fenbendazole, moxidectin, milbemycin oxime, fecal egg count reduction test, egg hatch assay, β-tubulin, Parasites & Vectors