S. aureus in Sheep Milk After Antibiotic Treatment: What the Withdrawal Period Doesn’t Fix

The antibiotic course is done. The sheep passed the withdrawal period. The milk is back on the market. So why are we still talking about Staphylococcus aureus?

Ewe’s milk makes up less than 1% of global dairy output, but in places like Greece, Spain, and southern Italy, it underpins entire regional food cultures — artisanal cheeses, traditional ferments, small-farm economies that go back centuries. It is also, unfortunately, where Staphylococcus aureus tends to make itself at home.

Mastitis caused by S. aureus is a persistent headache for small ruminant producers. The bacterium colonizes teat and skin surfaces, embeds itself in the mammary gland, forms biofilms, and sheds intermittently into milk. Veterinarians treat it with antibiotics. The standard approach is to use a withdrawal period — a set number of days after the last dose — before milk re-enters the food chain. After that, legally, the milk is safe to sell.

A recent study from the University of Ioannina looked at what actually happens to S. aureus populations in raw ewe milk after three commonly used treatment protocols. What they found should give dairy producers and food safety professionals something to think about.

What the study did

Researchers enrolled 75 sheep farms in the Epirus region of northwestern Greece. They collected milk from three populations: clinically healthy animals, sheep diagnosed with clinical mastitis before treatment, and those same mastitis-affected sheep at the first eligible milking after withdrawal.

The three treatment protocols were:

• Protocol 1: Penicillin G + dihydrostreptomycin (100 animals)
• Protocol 2: Oxytetracycline (50 animals)
• Protocol 3: Enrofloxacin (50 animals)

In total, 500 milk samples were analyzed. For each treated animal, the researchers had a pre-treatment sample and a post-withdrawal sample from the same individual — a paired design that allowed direct within-animal comparison. They tracked S. aureus culture outcomes, antimicrobial resistance profiles across 20 antibiotics, mecA gene carriage, MRSA markers, enterotoxin production, and biofilm adherence.

The bacteria stayed

The first and perhaps most striking finding: S. aureus detection rates did not drop meaningfully after any of the three treatments.

Detection in healthy control animals ran at 23%. Pre-treatment mastitis groups ranged from 22–30%. Post-withdrawal groups came in at 20–22%. Paired statistical analyses found no significant shift within any protocol (all p > 0.05). The bacteria were still there.

This should not surprise anyone working in this area. S. aureus in sheep mastitis systems behaves more like an endemic reservoir pathogen than a straightforward treatable infection. It sheds intermittently. It hides in biofilms. It can survive at the teat canal for hours post-milking and re-enter the gland. Even when a treated animal clears the infection temporarily, reinfection from the farm environment or other animals can re-establish colonization.

The study authors are careful here: they cannot confirm that the same strain persisted. Post-withdrawal positivity could reflect strain continuity, previously undetectable low-level colonization that rebounded after competitive pressure was removed, or fresh reinfection. The data cannot distinguish between these. But regardless of mechanism, the practical consequence is the same: withdrawal compliance did not reliably eliminate S. aureus from milk.

What antibiotics may actually change

Culture rates did not shift. Resistance profiles, however, showed a more complicated picture.

For beta-lactams, the situation was stable across all protocols. No new resistance acquisitions, no significant resolution events. Beta-lactam resistance was already very high before treatment — 98% of pre-treatment mastitis isolates showed beta-lactam resistance class-wide — and stayed that way. There was nothing for the antibiotics to change because the baseline was already near the ceiling.

The more interesting story is in the enrofloxacin group. Post-withdrawal isolates from Protocol 3 showed a pronounced jump in fluoroquinolone resistance: enrofloxacin resistance went from essentially absent in pre-treatment isolates to 57.1% in post-withdrawal isolates; ciprofloxacin resistance followed a similar pattern, reaching 42.9%. The Protocol 3 post-withdrawal group also showed the highest overall antimicrobial resistance burden in the dataset, with a weighted MAR index of 0.277 — roughly double the pre-treatment figures.

This pattern is consistent with how fluoroquinolones work. They are bactericidal and have broad activity, but resistance can emerge or be selected for relatively quickly. Strains that carry pre-existing resistance determinants, even at low frequencies, may survive exposure and come to dominate the post-treatment population.

The beta-lactam picture deserves separate attention. The mecA gene, which encodes PBP2a-mediated methicillin resistance and is the defining molecular marker of MRSA, was present across all groups. In controls (Group A), mecA was detected in 13.6% of isolates. In post-withdrawal Group C1 (penicillin-treated), that figure rose to 54.5%, the highest in the study. Composite MRSA-confirmed classifications showed that methicillin-resistant strains were present in every experimental group, including clinically healthy controls. In paired animals, mecA status was predominantly stable between pre- and post-treatment sampling. Where it changed, it generally came with concordant phenotypic markers.

MRSA-compatible strains were already circulating at the flock level before any treatment began. Short-term antibiotic courses did not dramatically shift that baseline, but they also did not eliminate it.

Toxin production: the part that gets less attention

Antimicrobial resistance in dairy systems tends to get the most coverage in food safety conversations. But this study’s findings on enterotoxin production are arguably more immediately relevant to consumer risk.

S. aureus strains can produce several classical enterotoxins (SEA through SED) that cause staphylococcal food poisoning — the kind that sets in within a few hours of eating contaminated food and involves vomiting, nausea, and abdominal cramping. Crucially, these toxins are heat-stable. Pasteurization kills the bacteria; it does not necessarily destroy preformed toxins.

The data:

Before treatment, the dominant enterotoxin types in mastitis-associated isolates were SEC and SED. SEA and SEB were absent from all pre-treatment mastitis groups.

After treatment, the picture shifted. SEA and SEB appeared in post-withdrawal isolates. SED increased significantly in paired analysis (p = 0.002). Overall enterotoxigenicity in the enrofloxacin group (C3) reached 85.7% — six of seven isolates produced at least one enterotoxin. In the penicillin group (C1), 45.5% of post-withdrawal isolates were enterotoxigenic, up from 30% pre-treatment.

Because strain typing was not performed, the researchers cannot confirm whether these are the same strains acquiring new toxin production capacity or different strains taking their place. Either way, the isolates detectable in milk after the withdrawal period were more toxigenically active than the isolates present before treatment. The food safety implication does not change.

Biofilm: why chronic mastitis is so difficult to treat

The study also tracked biofilm adherence before and after treatment using a microtiter plate assay, with paired within-animal comparisons available for 19 animals.

In 17 of those 19 animals, biofilm adherence strengthened after treatment. None showed a decrease. The shift was from predominantly moderate adherence before treatment to predominantly strong adherence after treatment. Sign test p = 1.5 × 10⁻⁵; Wilcoxon signed-rank p = 2.0 × 10⁻⁴.

Biofilm formation is one of the main reasons S. aureus mastitis is so difficult to resolve. Bacteria embedded in biofilm matrices are substantially less susceptible to antibiotics than free-floating (planktonic) cells. If treatment selects for or reveals a post-treatment population enriched in strongly adherent strains, that helps explain why bacteriological cure rates in S. aureus mastitis are often disappointing even when in-vitro susceptibility testing suggests the antibiotic should work.

Multi-drug resistance across the board

Across the full dataset of 113 S. aureus isolates, 35 (31.0%) were classified as multi-drug resistant (resistant to five or more of the 20 tested antibiotics). MDR isolates showed up in every group: controls (22.7%), pre-treatment mastitis isolates (31.4%), and post-treatment groups. Group C3 had the highest resistance burden overall.

Some context helps. Greece contributes roughly 15% of European small ruminant milk production. The farms in this study are not outliers — they represent typical practice in a major dairy sheep region. The fact that a third of all recovered S. aureus isolates, including those from clinically healthy animals, show multi-drug resistance profiles is not a finding specific to sick flocks with heavy antibiotic use.

What this means in practice

The study does not say that antibiotics are useless for ovine mastitis. What it documents is that the first post-withdrawal milking may carry S. aureus populations with enhanced persistence traits, higher toxigenic potential, and in some cases elevated resistance profiles compared with pre-treatment isolates.

Here are a few things worth sitting with:

For producers: Bacteriological cure of S. aureus mastitis with standard intramammary therapy is genuinely difficult. Blind antibiotic treatment without prior susceptibility testing is likely wasting drugs and possibly selecting for more problematic strains. Culture-guided therapy is worth the additional effort.

For herd health management: The presence of mecA-positive MRSA-compatible isolates in clinically healthy controls suggests that MRSA is circulating at the flock level, not just in sick animals. Milking hygiene, segregation of chronic shedders, and transmission control matter as much as antibiotic choice.

For food safety: Withdrawal periods are a legal compliance mechanism, not a biological guarantee. Post-withdrawal milk from treated animals may still contain enterotoxigenic S. aureus strains with strong biofilm capacity. This matters most in raw milk cheese production.

For antimicrobial stewardship: Enrofloxacin was associated with the most pronounced post-treatment resistance changes in this dataset — fluoroquinolone resistance jumped from near zero to over 57% in the post-withdrawal group. Fluoroquinolones are critically important in both veterinary and human medicine, and their use in routine mastitis treatment warrants closer scrutiny.

The broader picture is not complicated: antibiotic practices on small farms connect to a food chain hazard that does not disappear when the withdrawal clock runs out. The bacteria persist. Their virulence profiles may shift. And the milk, legally compliant, moves into the production chain.

FAQ

Q: What is Staphylococcus aureus and why does it matter in sheep milk?

S. aureus is a gram-positive bacterium that commonly colonizes the skin and mucosal surfaces of both animals and humans. In dairy sheep, it is one of the most frequent causes of mastitis — infection of the mammary gland. Beyond its role in animal disease, it can produce heat-stable enterotoxins that cause food poisoning, and some strains carry antibiotic resistance genes including mecA (associated with MRSA). Because sheep are used to produce raw milk for artisanal cheeses and other minimally processed dairy products, contamination of the milk chain with toxigenic or resistant S. aureus strains is a genuine food safety concern.

Q: What were the three antibiotic treatments compared in this study?

The study compared three protocols routinely used in Greek sheep farms: (1) penicillin G combined with dihydrostreptomycin, (2) oxytetracycline, and (3) enrofloxacin. Each treated group was sampled before treatment and again at the first milking after the legal withdrawal period ended.

Q: Did the antibiotics reduce S. aureus levels in milk?

Not significantly. Detection rates before treatment ranged from 22–30% across mastitis groups, and post-withdrawal rates came in at 20–22%. Statistical analysis of paired pre- and post-treatment outcomes found no significant differences within any of the three protocols. A higher proportion of treated animals did clear the infection compared to those that remained positive, but new post-withdrawal detections also occurred in animals that had been culture-negative before treatment.

Q: What is mecA and why is it important?

mecA is a chromosomal gene that encodes a modified penicillin-binding protein (PBP2a), which reduces the binding affinity of beta-lactam antibiotics — the most widely used class in mastitis treatment. Strains carrying mecA are classified as methicillin-resistant S. aureus (MRSA). In this study, mecA was detected across all experimental groups, including healthy controls, suggesting it is already circulating at the flock level before any antibiotic is used.

Q: Why did enterotoxin profiles change after treatment?

Before treatment, the dominant enterotoxin types in mastitis-associated isolates were SEC and SED, and SEA and SEB were absent. After treatment, SEA and SEB appeared in post-withdrawal isolates, and SED increased significantly (p = 0.002). The most likely explanations are that antibiotic pressure selected for strains carrying these toxin genes, or that enterotoxigenic strains — previously present at low levels or masked by other bacteria — became more detectable once competitive strains were removed. The researchers note that strain typing was not performed, so the precise mechanism cannot be confirmed from this data alone.

Q: What is biofilm and why does it complicate mastitis treatment?

Biofilm is a structured community of bacteria embedded in a self-produced matrix of proteins and polysaccharides. Bacteria living in biofilms are substantially harder to kill with antibiotics than free-floating (planktonic) cells — estimates range from 10 to 1,000 times less susceptible, depending on the strain and antibiotic. S. aureus in the mammary gland often grows in biofilm mode, which helps explain why bacteriological cure rates are frequently disappointing even when in-vitro tests suggest susceptibility. This study found that post-withdrawal isolates showed significantly stronger biofilm adherence than pre-treatment isolates, which suggests that antibiotic exposure may have enriched for the more persistent phenotypes.

Q: Is enrofloxacin particularly problematic?

Of the three protocols studied, enrofloxacin (Protocol 3) was associated with the most pronounced post-treatment resistance changes. Fluoroquinolone resistance in the post-withdrawal group jumped from near zero to over 57% for enrofloxacin and 43% for ciprofloxacin. The post-withdrawal enrofloxacin group also showed the highest overall antibiotic resistance burden in the dataset and the highest rate of enterotoxin production (85.7% of isolates). Fluoroquinolones are classified as critically important antimicrobials for human medicine, so their use in routine livestock mastitis treatment warrants careful consideration.

Q: Does passing the withdrawal period guarantee milk safety?

Not entirely. Withdrawal periods are legally defined timeframes that allow antibiotic residues to drop below regulatory limits — they are not a biological guarantee that all pathogens have been eliminated. This study found that S. aureus was still detectable at similar rates in post-withdrawal milk, with shifts toward more enterotoxigenic and biofilm-forming phenotypes. Raw milk from treated animals that re-enters the production chain may still carry strains with meaningful food safety relevance.

Q: What should dairy farmers do differently based on this research?

Culture-guided antibiotic selection (rather than blind treatment) is a consistently supported recommendation. Milking hygiene, early identification of chronic shedders, and segregation protocols can limit within-flock transmission more reliably than repeated antibiotic courses. Given the MRSA findings in this study, biosecurity measures — including hygiene during milking and limiting animal movement — are also relevant. For producers making raw milk products, the toxin findings add weight to the case for robust food safety controls beyond withdrawal compliance alone.

Q: How does this connect to antibiotic resistance in humans?

S. aureus and MRSA can circulate between livestock and humans through direct contact, shared environments, and the food chain. Raw milk is a potential transmission route. mecA-carrying isolates from sheep milk that enter the food chain or contaminate farm workers’ environments could contribute to MRSA reservoirs at the animal–human interface. Several studies have documented genetic overlap between livestock-associated and human-associated MRSA strains, which is why surveillance and stewardship at the farm level have human health implications beyond the animals themselves.