Benefits
Caries risk reduction (WB21)
Mouth bacteria (WB21): in 8 volunteers with no control group, 2.0×10⁹ CFU/day for 2 weeks significantly decreased mutans streptococci (P=0.039). In the randomized comparison against TI 2711, Ovalgen DC and xylitol, mutans streptococci did not differ significantly between groups; what rose in the WB21 arm was lactobacilli. No trial has measured actual cavities, and the tablets were dissolved on the tongue rather than swallowed, so this says nothing about a swallowed capsule.
Periodontal health heat-killed CP3365
Periodontal health (heat-killed CP3365): in 64 healthy adults with no gum disease, 8 weeks of heat-killed CP3365 held plaque control record, bleeding on probing and probing pocket depth steady while all three worsened on placebo (P<0.05). The gaps were small, about 0.08 mm on pocket depth, and the trial was run and funded by the ingredient maker. 16S rRNA showed Porphyromonas, Fusobacterium, and Haemophilus relative abundance significantly decreased. These were heat-killed, non-viable cells, so the result belongs to a postbiotic tablet and does not transfer to a live L. salivarius supplement.
Stool bacteria change (UCC118)
Gut barrier (UCC118): fermented milk-borne L. salivarius UCC118 in healthy adults significantly increased fecal enterococci and lactobacilli, while bifidobacteria, coliforms, and bacteroides remained unchanged. Counting bacteria in stool does not measure the gut barrier, so this study does not show a barrier effect. A separate small crossover trial in seven healthy adults did measure gut permeability after hard exercise and found a change in one of three markers, which is a preliminary signal at best. Neither study is among the references below.
CECT5713 in healthy adults
CECT5713 has been tested in one small trial in 40 healthy adults taking capsules for 4 weeks. It shifted stool bacteria counts, and the authors also reported changes in some blood immune markers. That trial is not among the references on this page, and a stool bacteria count is not itself a measure of immune function, so nothing cited here backs an immune claim.
Ls-33 obesity / gut barrier
Ls-33 was tested in a 12-week randomized trial in 50 obese adolescents. It shifted some fecal bacterial ratios but did not change bacterial cell numbers or short-chain fatty acids, and a companion report from the same trial found no benefit on inflammatory markers or on measures of metabolic syndrome. No gut barrier measure was taken. The authors themselves concluded the shift was unrelated to metabolic syndrome, so this is a null result rather than support for an obesity or gut barrier use.
Strain-specific clinical evidence
L. salivarius shows distinct strain-specific evidence across indications — WB21 (oral health), CP3365 (periodontal heat-killed postbiotic), UCC118 (gut), CECT5713 (immune), Ls-33 (tested for metabolic effects in obese adolescents, results null), LS01/DSM 22775 (atopic dermatitis), LI01 (animal liver-injury studies only, no human trial), PS2 (studied for preventing mastitis in late pregnancy, not metabolic health). Generic L. salivarius supplements are not equivalent to evidence-based strains.
Surface protein gut adherence
L. salivarius carries surface proteins that help it stick to the gut lining in laboratory work. Whether it stays in the gut after you stop taking it has not been shown in people. Like most probiotics it is best assumed to be transient, so keep taking it if you want the effect.
Mechanism of action
Strain-specific bioactivity
Different L. salivarius strains exhibit distinct clinical evidence and bioactivity profiles. Foundation principle: L. salivarius is not monolithic — strain-specific selection determines clinical relevance for any given indication.
Surface protein gut adherence + colonization
Surface proteins help the bacteria stick to the gut lining in laboratory work. Whether that produces lasting colonization in people has not been shown.
Mutans streptococci reduction (oral)
WB21 and related oral strains are proposed to lower S. mutans, a bacterium linked to cavities, when a tablet is dissolved in the mouth. The randomized comparison did not show a significant fall between groups.
Lactic acid + bacteriocin production
Lactic acid production lowers local pH; bacteriocin production provides targeted antimicrobial activity. Multi-mechanism antimicrobial activity profile.
Anti-inflammatory immunomodulation (Ls-33)
Mouse studies of Ls-33 report gut barrier and anti-inflammatory effects. The human trial that tested this idea in obese adolescents found no benefit on inflammation or on metabolic measures, so the mouse mechanism did not carry over.
Postbiotic heat-killed activity (CP3365)
Heat-killed CP3365 retains clinical activity without live bacteria — postbiotic application. The trial enrolled healthy adults only, so there is no evidence about use in immunocompromised people. A non-living form may still suit people who prefer not to take live bacteria.
Clinical trials
WB21 randomized open-label trial plus short-term 8-volunteer administration (UMIN000013160). 2.0×10⁹ CFU/day for 2 weeks significantly decreased mutans streptococci (P=0.039).
Volunteers in a randomised open-label caries-risk trial, plus a short-term study in 8 volunteers. Registered as UMIN000013160.
WB21 randomized open-label trial plus short-term 8-volunteer administration (UMIN000013160). 2.0×10⁹ CFU/day for 2 weeks significantly decreased mutans streptococci (P=0.039). In that randomized open-label trial, 64 healthy adults in four groups dissolved tablets on the tongue; mutans streptococci did not differ significantly between groups, and what rose in the WB21 arm was lactobacilli. The P=0.039 fall came from a separate 8-person study with no control group. No cavity outcome was measured, and this trial is not among the references below.
Heat-killed CP3365 (UMIN000045656) double-blind clinical trial (n=64) × 8 weeks.
64 adults in a double-blind placebo-controlled trial of a heat-killed preparation. Registered as UMIN000045656.
Heat-killed CP3365 (UMIN000045656) double-blind clinical trial (n=64) × 8 weeks. 64 healthy adults with no gum disease were randomized and 54 completed. Over 8 weeks all three measures worsened on placebo and stayed flat on CP3365 (P<0.05), but the gaps were small: plaque record +5.99% versus +0.44%, pocket depth +0.10 mm versus +0.02 mm. Relative abundance of the genera Porphyromonas, Fusobacterium and Haemophilus fell, while P. gingivalis and F. nucleatum themselves did not change. The bacteria were heat-killed, the study was run by the ingredient maker, and it is not among the references below.
Ls-33 fecal microbiota study in obese adolescents.
Obese adolescents in a faecal microbiota study.
Ls-33 fecal microbiota study in obese adolescents. That trial randomized 50 obese adolescents to Ls-33 or placebo for 12 weeks. It shifted some fecal bacterial ratios but did not change bacterial cell numbers or short-chain fatty acids, and a companion report found no benefit on inflammatory markers or metabolic syndrome measures. No gut barrier measure was taken. This is a null result, not evidence for an obesity or gut barrier use, and it is not among the references below.