CTX-M group 1 in birds, goats and water: One Health findings

A One Health study in the Kenyan Rift Valley found CTX-M group 1 genes in 14 of 79 screened bacterial isolates from migratory Little Stints, goats and water. All 13 bird and goat carriers were sampled in the pre-departure window, which points to local acquisition at shared freshwater hubs rather than import from the Arctic. The one water carrier came from an artesian borehole, not from the lake itself.

Why this matters: ESBL-producing Enterobacterales beyond the clinic

ESBL-producing Enterobacterales sit on the World Health Organization list of critical priority pathogens. ESBL stands for extended-spectrum beta-lactamase, the enzyme class that inactivates third-generation cephalosporins such as cefotaxime, ceftazidime and ceftriaxone, the workhorse drugs for severe Gram-negative infections. Three enzyme lineages do most of this damage: CTX-M, TEM and SHV. Within the CTX-M family, group 1 dominates clinical, veterinary and environmental samples worldwide, and its flagship variant blaCTX-M-15 has spread from hospital clones into community and environmental settings. The gene itself appears to have escaped from the environmental genus Kluyvera, then rode high-risk lineages such as Escherichia coli ST131 and promiscuous plasmids across species and continents.

Once a hospital problem, ESBL epidemiology now runs through livestock, wildlife and water. Mobile genetic elements, especially conjugative plasmids, move these genes between hosts and between bacterial genera by horizontal gene transfer, and many ESBL producers simultaneously carry resistance to fluoroquinolones, aminoglycosides and macrolides. Two group 1 variants can also be maddeningly hard to tell apart: a well-known pair shares about 98.7 percent nucleotide identity and differs by a single amino acid, so conventional identification methods cannot cleanly separate them.

That difficulty matters for surveillance design. Migratory birds are increasingly read as sentinels and long-distance vectors: they forage in habitats contaminated with antimicrobial residues, pick up resistant bacteria, and carry them along flyways that connect heavily human-influenced stopovers with remote breeding grounds. Livestock, by contrast, act as stationary nodes. In intensive production systems, antimicrobial use and waste deposition drive local resistance; in extensive pastoral systems like the semi-arid Rift Valley, transmission instead runs through fluid ecological interfaces, where herds, wildlife and people share the same scarce water. Those shared freshwater points are where this study looked.

Study design: two lakes, two migratory windows, three hosts

Researchers ran a repeated cross-sectional One Health surveillance across two contrasting alkaline lakes in the Kenyan Rift Valley, sampling three host domains in two migratory windows. The first window caught Little Stints within days of arriving from the Arctic in October 2021; the second caught them again in April 2022, just before they left. Both lakes are hyper-alkaline, but they differ in almost everything else that could shape resistance.

Site	Setting	pH	Pressure profile
The protected reserve lake	National reserve, no industrial extraction, agropastoral surroundings	9.0 to 10.2	Low direct industrial pressure; a highly connected stopover for resident and migratory birds
The industrial mining lake	Trona mining, urbanizing shoreline, severe freshwater scarcity	11.2	High anthropogenic pressure; sparse pastoral settlement

In total the team collected 389 samples: 184 from Little Stints, 157 from free-ranging domestic goats, and 48 water samples spanning lakes, springs, artesian boreholes, marshes and inflowing rivers. Birds were caught in mist nets, ringed to prevent resampling, and released immediately after non-invasive fecal collection. Goat sampling ran alongside structured questionnaires covering antimicrobial use, herd management and water, sanitation and hygiene (WASH) conditions.

Fecal samples were enriched in buffered peptone water and plated on MacConkey agar; water samples were vacuum-filtered through a series of cellulose membranes down to 0.45 micrometer and enriched overnight. Four to five colonies per sample, chosen for distinct morphology and lactose-fermentation phenotypes, were identified by MALDI-TOF mass spectrometry. Every Enterobacterales isolate then went through Kirby-Bauer disk diffusion testing against 12 agents from eight antimicrobial classes, with phenotypic screening for ESBL by the double-disk synergy test (DDST) under CLSI guidance. Multidrug resistance (MDR) was scored as non-susceptibility to at least one agent in three or more antimicrobial categories.

Because sequencing is expensive exactly where it is needed most, the molecular tier ran on conventional PCR. A panel of 79 isolates, the 75 phenotypic ESBL producers plus four ESBL-negative Klebsiella pneumoniae controls, was screened for four beta-lactamase targets.

PCR target	Amplicon size	Annealing temperature
CTX-M group 1	255 bp	61 degrees Celsius
blaTEM	786 bp	50 degrees Celsius
blaSHV	356 bp	50 degrees Celsius
blaOXA-1-like	820 bp	62 degrees Celsius

From 389 samples to 14 carriers

Every step in the funnel changes the denominator. From 389 samples the team recovered 507 Gram-negative isolates, of which 479 belonged to the order Enterobacterales. Phenotypic screening confirmed 75 ESBL-positive Enterobacterales from 68 distinct samples (30 birds, 30 goats, 8 water), since some samples yielded more than one producer. Among those 75, the majority, 52 isolates or 69.3 percent, were also multidrug resistant.

Screening step	Isolates	Share
Gram-negative isolates recovered	507	Baseline
Enterobacterales	479	94.5 percent of Gram-negative isolates
Phenotypic ESBL-positive	75	15.7 percent of Enterobacterales
Selected for PCR screening	79	75 ESBL producers plus 4 controls
CTX-M group 1-positive	14	17.7 percent of isolates screened

PCR then found the gene group in 14 isolates: 13 phenotypic ESBL producers and, notably, one Klebsiella pneumoniae that had screened ESBL-negative. That last detail is the first surprise in this dataset, and it gets its own section below.

Which species carried CTX-M group 1

Among the 14 carriers, Escherichia coli was the most common species with 6 isolates, followed by Enterobacter cloacae with 4, Klebsiella pneumoniae with 3, and a single Providencia rettgeri.

Species	CTX-M group 1-positive isolates
Escherichia coli	6
Enterobacter cloacae	4
Klebsiella pneumoniae	3
Providencia rettgeri	1

The wider pool of 75 ESBL-positive Enterobacterales looked different. It was dominated by Enterobacter species (31 isolates, of which 27 were E. cloacae), followed by E. coli (17) and Klebsiella species (12, mostly K. pneumoniae). Lower-frequency finds included Leclercia adecarboxylata (6), Citrobacter freundii (6), and one isolate each of Raoultella ornithinolytica, Providencia rettgeri and Pantoea eucrina. The gene group is clearly not the property of one bacterial brand: it circulates across the order.

Where and when the detections clustered

Distribution across hosts and sites is where the ecology starts to show. Little Stints contributed 8 of the 14 carriers, goats 5, and water just 1. By site, the protected reserve lake accounted for 8 carriers and the industrial mining lake for 6.

Dimension	Category	Isolates
Host	Little Stints	8
Host	Goats	5
Host	Water	1
Site	Protected reserve lake	8
Site	Industrial mining lake	6
Window	Pre-departure, birds and goats	13
Window	Arctic arrival, water only	1
Phenotype	ESBL and multidrug resistant	11
Phenotype	ESBL, not multidrug resistant	2
Phenotype	ESBL-negative	1

Every single avian and livestock carrier came from the pre-departure window. Not one newly arrived Arctic bird carried the gene group. The lone water carrier came from a freshwater artesian borehole at the protected lake, sampled during the arrival window, and it was never found again in the pre-departure round.

Co-resistance and co-carriage: what traveled with the gene

The 14 carriers arrived loaded. Eleven of them were multidrug resistant. Resistance was near-universal to ampicillin (12 of 14) and common to tetracycline (10) and sulfamethoxazole-trimethoprim (6). Critically important antimicrobials were not spared: gentamicin (5), cefepime (5) and ciprofloxacin (2) all recorded non-susceptible isolates.

Co-resistance agent	Resistant isolates, of 14
Ampicillin	12
Tetracycline	10
Sulfamethoxazole-trimethoprim	6
Gentamicin	5
Cefepime	5
Ciprofloxacin	2

Co-carriage of other beta-lactamase genes was routine: 9 isolates carried blaTEM, 2 carried blaSHV, and 1 carried an OXA-1-like gene. The study did not sequence the mobilizing elements, but conjugative IncF plasmids, ISEcp1 insertion sequences and ISCR1 elements are well documented movers of CTX-M genes and MDR cassettes across hosts, which makes the observed stacking of resistance classes unsurprising.

Silent carriage: the ESBL-negative isolate that carried the gene group

The first counterintuitive finding is a quiet one. One Klebsiella pneumoniae isolate from a goat carried the gene group while screening completely ESBL-negative. Resistance genes can sit silent, unexpressed under standard culture conditions, and a phenotype-only surveillance protocol would have missed it entirely. The design of this study anticipated the possibility: the PCR tier deliberately reached beyond phenotypic ESBL producers to include a 50 percent subset of the ESBL-negative K. pneumoniae recovered, precisely because that species is a prolific host and shuttle for MDR plasmids. The discrepancy ran both ways, too: some isolates with a resistant phenotype lacked every targeted gene, a reminder that candidate-gene PCR panels have hard specificity limits. One silent carrier cannot define a population frequency, but it redraws the surveillance question from what bacteria do in a dish to what genes they actually carry.

The open lakes were sterile; the small freshwater hubs were not

The second surprise is spatial. Both study lakes are hyper-saline and hyper-alkaline, and neither lake water yielded the gene group. Every positive water sample came from a peripheral multi-use freshwater hub, in this case an artesian borehole, where livestock, wildlife and people converge on the same scarce drinking water. Extreme alkalinity and salinity make the lakes themselves inhospitable to most Enterobacterales, so the risk is not distributed across the water body at all: it is concentrated in the small, shared, biologically dense points around its edge.

The drier lake had zero detections

The third surprise inverts a common assumption. The industrial mining lake suffers the region’s most severe freshwater scarcity, which ought to concentrate resistance. Yet it produced no positive water samples, while the better-watered protected lake produced the only one. Scarcity alone is not the driver. What the protected lake has instead is connectivity: as a major stopover it pulls in a broader, more heterogeneous mix of resident and migratory birds, and that seasonal influx can deliver a pulse of resistance elements into the peripheral hubs. Water limitation without the right host traffic buys nothing.

Birds arrived clean and left carrying: the migratory baggage pattern

The fourth surprise is temporal, and it reframes the birds from couriers to students. All 13 avian and livestock carriers were sampled in the pre-departure window, in the months following the long dry season, when shrinking water points force dense crowding of animals and people around the few remaining hubs. The same window is when shorebirds enter hyperphagia, dramatically raising water and food intake to fuel the northward flight, which plausibly accelerates both exposure and gene acquisition at those bottlenecks. Add months of prior foraging at human-influenced stopovers along the flyway, and the migratory baggage hypothesis writes itself: birds pick up resistance locally during the stay and carry it onward, rather than importing it from the Arctic. The arrival-window pattern is the mirror image: birds came in during the short rains, when rain expands and dilutes contaminated water points, from isolated high-latitude breeding grounds with minimal human contact.

What this means for surveillance programs

• Sample phenotype-blind. Include ESBL-negative Klebsiella and other priority genera in PCR panels; phenotype-only screening misses carriers.
• Sample the hubs, not the water body. Boreholes, springs and marshes are where resistance concentrates; open hyper-alkaline water is close to uninformative.
• Time the windows. Dry-season and pre-departure periods concentrate both hosts and genes; arrival-window sampling after rains will read low.
• Tier the methods. Conventional PCR is a scalable first tier for resource-limited settings; reserve whole-genome sequencing for resolving lineages, plasmids and expression.
• Write wildlife and environment into national AMR action plans. Passive clinical monitoring misses the interface where flyways meet pastoral water.
• Protect the water points. Upgrading WASH infrastructure at communal hubs attacks the selection pressure where host populations collide.

Limitations

The study is honest about its edges, and so should readers be. Without whole-genome sequencing, clonal lineages could not be mapped and plasmid-level architecture stayed unresolved; no long-read or amplicon sequencing was performed. No environmental residue measurements were taken, so local selective pressures are inferred rather than quantified. There was no formal statistical testing, so risk factors are not estimated, and without satellite telemetry the actual flyways and stopover histories of the sampled birds remain assumptions. The PCR panel can only find the genes it targets, and the single silent-carrier observation is exactly that, a single observation. Two group 1 variants in this family also share about 98.7 percent nucleotide identity, which limits what panel-level PCR can say about which variant is which.

Bottom line

In a flyway where global migration meets scarce water, CTX-M group 1 pools at shared freshwater hubs, in birds and goats that acquire it locally during the dry-season stopover, and in carriers that phenotype screens cannot see. The Rift Valley reads less like a passive corridor and more like a dynamic epicenter where global flyways and local land use together tune the resistance signal. Surveillance that follows the water, the season and the genes, not just the clinic, is the one that will see it coming.

FAQ

What is CTX-M group 1?

CTX-M group 1 is the dominant subgroup of CTX-M extended-spectrum beta-lactamases, enzymes that inactivate third-generation cephalosporins such as cefotaxime. Its best-known member, blaCTX-M-15, is the globally dominant variant across clinical and environmental samples. Group 1 genes turn up in hospital, community, livestock and environmental settings, making them a useful cross-sector marker.

What are ESBL-producing Enterobacterales?

ESBL-producing Enterobacterales are order-level Gram-negative bacteria, including Escherichia coli, Klebsiella pneumoniae and Enterobacter cloacae, that make enzymes able to destroy extended-spectrum beta-lactams. The World Health Organization lists them as critical priority pathogens. Treatment options narrow once third-generation cephalosporins fail, which is why they anchor AMR surveillance.

How do migratory birds spread antimicrobial resistance?

Migratory birds pick up resistant bacteria while foraging in habitats contaminated with antimicrobial residues, then carry them along flyways that span continents. Those flyways link heavily human-influenced stopovers with remote breeding grounds, letting birds bridge contaminated and isolated environments. In this study the gene group turned up in shorebirds, goats and water at the same lakes.

What is the double-disk synergy test?

The double-disk synergy test, or DDST, is a phenotypic screen for ESBL production. A disk containing amoxicillin-clavulanic acid is placed near disks of third-generation cephalosporins on an agar plate inoculated with the isolate. Enhanced inhibition near the inhibitor disk indicates ESBL production, and the test is interpreted under CLSI standards.

Can bacteria carry resistance genes without testing resistant?

Yes. One Klebsiella pneumoniae isolate from a goat carried the gene group while screening ESBL-negative, meaning the determinant was phenotypically unexpressed under standard culture conditions. A phenotype-only surveillance protocol would have recorded that animal as clean. This is why PCR-based screening that reaches beyond resistant phenotypes adds real detection power.

Why do shared freshwater hubs matter for antimicrobial resistance?

Shared freshwater hubs force livestock, wildlife and people into the same small volume of water, concentrating hosts and antimicrobial residues together. Sub-inhibitory residues can co-select for resistant strains, and biofilms accelerate horizontal gene transfer between strains. In this study every positive water sample came from a freshwater hub, never from the hyper-alkaline lake water.

What is One Health surveillance?

One Health surveillance monitors resistance across human, animal and environmental domains as a single connected system rather than three separate ones. This study is an example: birds, goats and water at the same lakes were sampled and tested with the same methods. The approach is built for settings where sequencing is out of reach, since PCR can deliver scalable, low-cost coverage.

What is migratory baggage in AMR research?

Migratory baggage is the hypothesis that birds acquire resistance determinants locally during stopovers and then carry them onward to the next site or breeding ground. It predicts that detections should build up over the stay rather than appear on arrival. This study fits that pattern, since every avian and livestock carrier was sampled in the pre-departure window and none in the arrival window.

Did the study find CTX-M group 1 in water?

Yes, but only once and only at the edges. The single water carrier came from a freshwater artesian borehole at the protected lake during the arrival window, and it was absent from the same site in the pre-departure round. Neither hyper-saline lake body yielded the gene group, so open water was effectively sterile for these bacteria.

Which antibiotics were the isolates resistant to?

Among the 14 carriers, resistance was highest to ampicillin at 12 isolates, followed by tetracycline at 10 and sulfamethoxazole-trimethoprim at 6. Critically important agents were affected too: gentamicin and cefepime each had 5 non-susceptible isolates and ciprofloxacin had 2. Eleven of the 14 carriers met the definition of multidrug resistant.

Is PCR or sequencing better for AMR surveillance?

Neither alone. Conventional PCR is cheap, fast and scalable, which makes it a practical first tier where sequencing budgets do not exist, but it only finds the genes its panel targets. Whole-genome sequencing resolves clonal lineages, plasmid architecture and gene expression, and catches uncharacterized determinants. The efficient design runs PCR first and escalates anomalies to sequencing.

Can antimicrobial resistance spread between wildlife and livestock?

In principle yes, and shared water is the obvious meeting point, since biofilm structure brings veterinary and environmental strains into close physical contact where horizontal gene transfer can occur. This study found the same gene group in shorebirds, goats and a water hub at the same lakes, which is consistent with that exchange. Direction and mechanism still need sequencing to confirm.