Benefits
Upper respiratory symptoms in active adults and office workers
In a 16-week placebo-controlled trial in active adults, fewer people on the strain had a week with upper respiratory symptoms (66% vs 90%). Two winter trials in office workers run by Yakult staff found fewer symptoms than on a control milk, one counting symptom days rather than people. Trials in 243 athletes, 737 nursing home residents and 154 older day care users found no reduction.
Bowel habits and stool consistency in adults with hard stools
In 70 adults with chronic constipation symptoms, a daily 65 mL drink for 4 weeks improved self-rated severity and stool consistency more than placebo. A 40-woman pilot after childbirth improved symptom scores but not stool frequency, while trials in 90 adults with functional constipation and 64 healthy adults with hard stools missed their main outcomes.
Abdominal comfort in students during exam stress
In a placebo-controlled trial in 47 healthy medical students drinking the strain daily for 8 weeks before a national exam, an abdominal symptom score, the trial's main outcome, stayed lower than on placebo, and the rise in self-rated stress was smaller. The trial was run by Yakult researchers with a university team.
Sleep quality ratings in students during exam stress
In 94 medical students over two exam years, a daily 100 mL drink for 8 weeks before and 3 weeks after the exam gave better ratings on two sleep inventory factors, sleepiness on rising and sleep length, than placebo milk, and kept deep sleep and sleep onset time steadier on overnight recordings. The abstract does not name a primary outcome.
Abdominal symptom scores in adults with chronic bowel complaints
In 80 adults with irritable bowel syndrome, the strain twice daily for 8 weeks did not reach the main goal of a 30% drop in a combined symptom score, and the difference from placebo 8 weeks later was not significant. Funding is not stated in the abstract.
Mechanism of action
Cell wall polysaccharide and survival in acid and bile (lab work)
The strain carries a large cell wall polysaccharide, LCPS-1, whose structure Yakult researchers have mapped. In the lab, a mutant lacking it grew less in milk and survived acid and bile worse than the normal strain (19% vs 73% after both). This is lab work and does not show survival in people.
Immune cell markers in office worker and athlete trials
In office workers, the fall in natural killer cell activity seen on control milk did not occur on the strain, and a later trial found higher HLA-DR on dendritic cells and monocytes. In active adults, saliva IgA was higher than on placebo at weeks 8 and 16. These are markers, not illness outcomes.
Stress hormone and gut bacteria markers under exam stress
In medical students before an exam, saliva cortisol rose only in the placebo group, and the strain group kept a higher number of gut bacterial species with a lower share of Bacteroidaceae. These are markers measured in one maker-run trial and do not by themselves show a felt benefit.
Clinical trials
Randomized, double-blind, placebo-controlled trial; a daily L. casei Shirota (LcS) drink or placebo for 16 weeks of winter training, with weekly illness logs (Gleeson et al. 2011, Int J Sport Nutr Exerc Metab).
84 highly active men and women in endurance-based activities (42 per group); 58 completed (32 probiotic, 26 placebo).
The stated purpose was upper respiratory infection incidence: 66% on LcS vs 90% on placebo had at least one week with symptoms (p = 0.021), and episodes were fewer (1.2 vs 2.1, p < 0.01). Severity and duration did not differ. Saliva IgA was higher on LcS. The abstract does not state the dose or funding. Only 26 of 42 on placebo and 32 of 42 on LcS completed.
Randomized controlled trial; one bottle a day of fermented milk with at least 100 billion live LcS (YIT 9029) or plain milk as control for 12 winter weeks; the paper does not describe blinding (Shida et al. 2017, Eur J Nutr).
96 healthy male office workers aged 30 to 49 in Tokyo (49 LcS, 47 control).
The primary outcome, upper respiratory infection incidence, was 22.4% on LcS vs 53.2% on control milk (P = 0.002), with fewer symptom days and shorter episodes. The fall in natural killer cell activity seen on control milk did not occur on LcS. No adverse effects. All authors were Yakult Honsha employees, and the study had no outside grant.
Randomized, double-blind, parallel-group trial; one 80 mL bottle a day of fermented milk with 40 billion CFU LcS or a matched unfermented control milk for 28 days in January and February 2024; registered UMIN000052964 (Naito et al. 2025, Biosci Microbiota Food Health).
200 healthy office workers aged 23 to 59 in Tokyo (100 per group).
For the primary outcome, the share of days with upper respiratory symptoms over 28 days was 23.0% vs 33.3% on control (p = 0.022), with the gap seen in days 15 to 28, but cumulative incidence, the share of people with any symptoms, did not differ over the whole period (p = 0.150). Severity, a secondary outcome, was lower. HLA-DR on dendritic cells and monocytes rose. All authors were Yakult Honsha employees, with no outside grant.
Single-center, randomized, double-blind, placebo-controlled trial; one 65 mL pot (at least 6.5 billion live LcS) or matched placebo twice a day for 20 weeks (Gleeson et al. 2016, Eur J Appl Physiol).
243 university athletes and games players who completed the study (126 LcS, 117 placebo).
Upper respiratory symptom episodes were not reduced on LcS, and their duration and severity did not change; overall episode incidence was unexpectedly low (0.6 per person). Antibody titers to cytomegalovirus and Epstein-Barr virus fell in carriers on LcS, a marker only. The full text states that Yakult Honsha fully sponsored the study.
Randomized, double-blind, placebo-controlled trial in 53 nursing homes; two bottles a day of fermented milk (at least 6.5 billion live LcS each) or a drink without bacteria for 176 days, with flu vaccination after 21 days; registered NCT00849277 (Van Puyenbroeck et al. 2012, Am J Clin Nutr).
737 healthy adults aged 65 and over in Antwerp, Belgium (375 LcS, 362 placebo).
No effect on the primary outcomes: days with respiratory symptoms, the chance of respiratory symptoms (OR 0.87, 95% CI 0.62 to 1.29) or flu vaccine antibody response. The authors concluded the drink had no statistically or clinically significant protective effect. Funding is not stated in the abstract.
Randomized, double-blind, placebo-controlled trial; 65 mL a day of an LcS drink or a sensorially identical placebo for 4 weeks, with weekly questionnaires and medical exams (Koebnick et al. 2003, Can J Gastroenterol).
70 adults with symptoms of chronic constipation in Germany.
Self-reported constipation severity and stool consistency improved more on LcS from the second week (P < 0.0001), and 89% vs 56% on placebo rated their drink as helpful (P = 0.003). Flatulence and bloating did not change. No adverse reactions. The abstract does not give the cell count or funding.
Randomized, double-blind, placebo-controlled trial at two Belgian sites; one 65 mL bottle a day with at least 6.5 billion CFU LcS or matched non-fermented milk for 8 weeks; stool hardness measured with a texture analyzer (Tsujibe et al. 2024, Nutrients).
64 healthy adults, mostly women, with hard or lumpy stools on at least half of bowel movements during screening (32 per group).
The primary outcome, measured stool hardness, did not differ significantly (p = 0.052), and the share of normal stools did not differ. Stool water content was higher on LcS, and stools were softer in a subgroup with the hardest stools at baseline. The authors stated they could not show a softening effect. Yakult Honsha funded the study, and all six authors were Yakult employees.
Double-blind, placebo-controlled trial; one 100 mL bottle a day of fermented milk with more than 100 billion CFU LcS or a placebo milk matched for nutrients, color, flavor and taste for 8 weeks up to the day before a national exam, with symptom scores, saliva cortisol, blood gene expression and stool bacteria (Kato-Kataoka et al. 2016, Appl Environ Microbiol).
47 healthy medical students in Japan (23 LcS, 24 placebo).
The full text names the total abdominal dysfunction score as the primary endpoint; it was lower on LcS than on placebo over the trial (P < 0.05). Stress-related rises in a visual analog stress rating were also smaller, and saliva cortisol rose before the exam only on placebo. The LcS group kept more gut bacterial species. Yakult Honsha sponsored the study and supplied the drinks, and most authors worked for Yakult.
Multicenter, randomized, double-blind, placebo-controlled parallel trial; fermented milk with LcS or a placebo drink daily; the abstract does not give the dose or length (Fujita et al. 2013, Am J Infect Control).
154 older adults using 4 day care facilities in Tokyo (analyzed).
The number of people with an acute upper respiratory infection was almost identical (31 on LcS vs 32 on placebo), and infection events and symptom scores per day observed did not differ (P = 0.89 and 0.64); raw event counts were higher on LcS (68 vs 51). Mean duration per infection was shorter on LcS (3.71 vs 5.40 days). The abstract does not name a primary outcome or funding.