Streptococcus salivarius: a helpful microbe with a rare dark side

Ask people to name a bacterium that lives in their mouth and, if they can name one at all, they name the one that causes cavities. Almost nobody names the one that arrives first. In a small longitudinal study, swabs were taken from 12 healthy newborns on their first day of life and again at one, three, six, nine and twelve months; Streptococcus salivarius was the most frequently recovered organism in the mouth ¹. Before the first tooth, before the first meal, it is already there. Now — that was 12 babies ¹ in a single study published in Chinese, so hold the number loosely. But the picture it paints matches everything else we know about where this organism likes to sit.

What it is

Streptococcus salivarius is a Gram-positive, chain-forming viridans-group streptococcus of the family Streptococcaceae that lives as a normal commensal on the human tongue and in saliva. That is the whole definition. It is not a pathogen we tolerate; it is a resident that occasionally causes trouble when something puts it where it does not belong.

Where it lives

Humans are its only natural host. No environmental reservoir, no animal cycle — us, and nothing else ². Its preferred address is the dorsum of the tongue, and from there it spreads into saliva and across the soft tissues of the mouth and throat.

We have an unusually precise map of that, because when researchers dosed a volunteer with the probiotic strain K12 and then sampled 16 oral sites repeatedly by quantitative PCR, they could watch where it went ⁵. The tongue carried the highest counts. Saliva and the buccal membranes followed. Both sides of the pharynx were positive, though asymmetrically. And four samples from gingival sulci and four from tooth surfaces stayed negative at every time point ⁵. This is a soft-tissue and saliva organism, not a plaque organism — which matters, because the ecological company it keeps is not the company that drills holes in teeth.

What it does for us

Three mechanisms are worth knowing, in order of how well they are established.

It makes antibiotics. Not metaphorically. Some S. salivarius strains carry transmissible megaplasmid DNA encoding arrays of bacteriocins — targeted protein antibiotics — including the lantibiotics salivaricin A and salivaricin B ²³. The prototype probiotic strain, K12, was the endpoint of a roughly 40-year search through the genus Streptococcus for a harmless natural antagonist of Streptococcus pyogenes, the cause of strep throat ³. The chemistry is still being extended: a further K12 antimicrobial, salivabactin, was identified in 2024 ¹⁷, and a 2025 study showed that its production is switched on by environmental acidification — the organism arms itself precisely when the neighbourhood turns acidic ⁴.

It raises pH. S. salivarius is urease-positive: it hydrolyses urea to ammonia, and ammonia buffers acid ¹⁰. Alkali generation by oral bacteria — through urease and through the arginine deiminase system — is believed to be one of the ecological brakes on the initiation and progression of dental caries ⁹. Note the honest verb: believed to be. That is a well-argued mechanism with real supporting data, not a settled clinical fact.

It occupies space. Because it is numerically prominent on the tongue and armed with bacteriocins, S. salivarius has been proposed to act as a population surveillance and modulation system inside the oral microbiota ². In the dish, K12 suppresses the growth of Gram-positive bacteria implicated in oral malodour, among them Solobacterium moorei, Atopobium parvulum and Eubacterium sulci ¹⁶.

Which brings us to what the human trials actually show — and here I have to slow down, because this is where enthusiasm usually outruns the data.

For bad breath: 23 subjects ⁶ rinsed with chlorhexidine, then took K12 or placebo lozenges. One week in, 85 % of the K12 group and 30 % of the placebo group had substantial drops in volatile sulphur compounds ⁶. Encouraging. But a later double-blind, randomised, placebo-controlled trial in 28 people with tongue-coating halitosis — no tongue scraping, no mouthwash beforehand — found no significant difference between K12 and placebo on any measure ⁷. Read those two together and the conclusion is neither "it works" nor "it doesn't." It is: the coating has to come off first, and the probiotic is at best an adjunct.

For sore throat: a 2019 systematic review ⁸ pooled every randomised trial of K12 it could find. Four trials, 1 846 participants — and all four were judged poor quality on the Cochrane risk-of-bias assessment ⁸. One prophylaxis trial in children reported streptococcal pharyngitis in 16.2 % of the K12 group versus 48.6 % of controls; another, placebo-controlled and larger, found 7.8 % versus 8.8 %, which is nothing ⁸. Given to adults alongside penicillin, K12 showed no significant benefit ⁸. The reviewers' verdict: safe, well tolerated, role not yet established ⁸. That is the honest state of the evidence, and no amount of packaging changes it.

And whatever a strain does, it does not stay long. In that same PCR tracking study, K12 remained detectable on the oral mucosa for as long as three weeks after dosing, but counts fell steadily after day 8 ⁵. Colonisation is a tenancy, not a purchase.

What it does against us

Here is the part that gets left out of the marketing.

S. salivarius is poorly equipped to invade — it has a scarcity of virulence attributes and rarely ventures far from the tongue in a healthy host ². But "rarely" is not "never," and the exceptions share a pattern: they happen when a needle, a scope or a damaged mucosa carries the organism past the barrier.

The clearest example is iatrogenic meningitis. A review of 60 cases of meningitis following lumbar puncture found streptococcal species responsible in 33 of the 52 cases that yielded an isolate — 63 % — with a median incubation of 24 hours, and S. salivarius among the named culprits ¹¹. The mechanism is unglamorous and entirely preventable: droplets from the operator's mouth. One hospital reported 6 cases of meningitis after spinal anaesthesia over 5 years, all associated with a single anaesthetist and, the authors judged, with possible breaches of aseptic technique; S. salivarius was confirmed in three of them — one by culture, two by 16S rRNA gene sequencing ¹².

It has also been reported in infective endocarditis, including a fatal case complicated by mycotic aneurysms, reported precisely because the species is so routinely written off as a blood-culture contaminant ¹³. And in patients with cancer, viridans-group streptococci become genuine bloodstream pathogens: in one four-year hospital series of 43 such patients, S. salivarius accounted for 2 isolates — 4.7 % — with mortality across the whole group of 18.6 % ¹⁴.

None of this makes S. salivarius a villain. It makes it a commensal, which is an organism defined by its address. On the tongue it is a good neighbour. In the cerebrospinal fluid it is a medical emergency.

The name, old and new

The current, correct name is Streptococcus salivarius Andrewes and Horder 1906, retained on the Approved Lists in 1980 ¹⁵. Streptococcus salivarius subsp. salivarius is a synonym of the species itself.

The confusing part is the yoghurt organism. In 1984, on the strength of DNA–DNA hybridisation, Streptococcus thermophilus was reclassified as S. salivarius subsp. thermophilus; in 1991 the separate species was revived; and both names then circulated at once in research, industry and regulation ¹⁵. A 2025 study settled it with genomics — a core-gene phylogeny across 216 Streptococcus genomes, plus average nucleotide identity and digital DNA–DNA hybridisation — and concluded that S. salivarius and S. thermophilus are distinct species ¹⁵. So S. salivarius subsp. thermophilus is the obsolete name of a different organism, not an old name for the tongue resident.

What we still don't know

We do not know whether deliberately raising S. salivarius reduces caries — tooth decay — in humans: the alkali mechanism, ammonia buffering plaque acid, is plausible, but the clinical trials have not been done ⁹. We do not have a single large, low-risk-of-bias randomised trial of the probiotic strain K12 for sore throat — one built so the result cannot be tilted ⁸. And we do not know what dose or schedule keeps a strain in the mouth, only that a short course fades within weeks ⁵. Anyone telling you more than that is telling you more than the literature says.