There is no single recommended dose of Lactobacillus paracasei, because the trials that produced the results were run on different strains at doses the abstracts often do not report. The best-studied strain for cholesterol, TISTR 2593, was given daily for 90 days to 50 adults and lowered LDL cholesterol significantly — but the published abstract does not state the CFU count. A dose question about “L. paracasei” in general cannot be answered, because L. paracasei is a species, not a product.
How much L. paracasei should you take?
No single figure can be given, and a label that prints one is describing its own product rather than a dose established by trials. The reason is strain specificity: a result obtained with TISTR 2593 does not transfer to DKGF1, to Shirota, or to a generic “L. paracasei” on a shelf. Each strain is a different organism with different adhesion, different metabolite output, and different trial history.
What can be said is the shape of the evidence. The cholesterol trial that reported a significant LDL reduction enrolled 50 adults with hypercholesterolaemia, randomised them to the strain or a maltodextrin placebo, and measured biomarkers at days 0, 45 and 90. Three timepoints over three months is the duration the result rests on. The CFU count is not in the abstract, which is a gap in the reporting, not a gap you can fill by assuming a standard number.
Does the CFU count on a label predict anything?
Not by itself, and the number is one of the least informative figures on a probiotic label. Three reasons, in order of how much they cost a buyer.
The count is at manufacture, not at expiry. A capsule declaring 10 billion CFU at the point of manufacture may hold far fewer by the time you take it. Only a label that states a count “through end of shelf life” is making a claim about what you actually swallow.
The count says nothing about strain identity. Ten billion CFU of a strain with no published trial is ten billion CFU of an untested organism. The genus and species on the front of the box do not tell you which strain is inside, and the strain is what the evidence is about.
Higher counts have not been shown to work better. The dose-response question for probiotics is largely unstudied; trials select a dose and report whether it worked, not whether a larger dose would have worked better. A 50-billion product is not evidence of a better product, only of a larger number on the box.
Why is the dose so hard to find?
Three structural reasons, and none of them is an oversight by the researchers.
Probiotic trials are run on products, not on ingredients. A trial buys or is given a finished capsule, measures the CFU at the start of the study, and reports outcomes. The strain, the dose, the capsule matrix and the storage conditions are all part of what was tested. When different laboratories use different preparations of what the label calls the same strain, the CFU numbers stop being comparable even when they are reported.
The number on the label is a specification, not a finding. A manufacturer sets a CFU target for a batch. No trial established that target as the dose at which the effect appears. The two numbers come from different processes — one from quality control, one from a clinical study — and they are routinely conflated in marketing.
Strain-level evidence is sparse by design. Running a trial on a specific strain is expensive, and the result is only commercially useful to whoever owns that strain. This is why the literature clusters: a few strains have several trials, most have none, and the ones with none still appear on labels. The absence of a dose figure for “L. paracasei” is not a gap that will be filled by better searching; it reflects how the research is funded.
The practical consequence is that a dosage question about a species cannot be answered, only a dosage question about a strain can. If a label names the strain and the strain has trials, look up what those trials used for dose and duration. If the label names only the species, there is nothing to look up.
Four outcomes moved and five did not, which is the most useful thing in the whole trial.
| Outcome | Direction | Significance |
|---|---|---|
| LDL cholesterol | lower | p < 0.05 |
| Malondialdehyde (oxidative stress marker) | lower | p < 0.001 |
| TNF-α (inflammatory marker) | lower | p < 0.01 |
| Apolipoprotein E | higher | p < 0.01 |
| Adiponectin | higher | p < 0.001 |
| Total cholesterol | no change | — |
| HDL cholesterol | no change | — |
| Triglycerides | no change | — |
| Total bile acids | no change | — |
The pattern matters. LDL fell while total cholesterol did not, which means the LDL fraction moved without the whole number following it. Triglycerides and HDL were flat. If you are buying a probiotic for a lipid panel, the outcome you would want to see move — total cholesterol or triglycerides — is the one that did not.
How should the dose be scheduled?
Daily, with the trial’s own schedule as the only anchor: once a day for 90 days, measured at 45 and 90. There is no published comparison of morning versus evening dosing for this strain, and no washout data that would tell you how quickly the effect reverses. Splitting the dose is not supported or contradicted by anything in the trial record; it is simply untested.
Ninety days is also the honest answer to “how long before it works”. The 45-day measurement is the earliest point at which the trial looked, so any claim of a faster effect comes from somewhere other than this study.
Is L. paracasei worth taking?
That depends on which outcome you want, and the evidence is narrower than the marketing suggests. Two of our earlier pages cover the parts of this question the dose cannot answer.
The L. paracasei evidence page runs all four metabolic outcomes with their confidence intervals, including the immune-result rows where the effect was real but small and the cumulative-risk row where the confidence interval crossed 1. The ingredient page lists what the species is and which strains have trial records at all.
If you are comparing finished products rather than strains, Gut Health Supplements Compared applies the same strain-and-dose check across labels, and the GutSave review runs it on a bottle that publishes a blend total instead of per-strain counts. For the general version of this problem — why a label can name ten ingredients and give one weight — the proprietary blend explainer sets out the labelling rule behind it.
FAQ
How many CFU of L. paracasei should I take? No trial-based number can be given for the species, because the strains tested differ and the abstracts do not always report CFU. The one cholesterol trial that produced a significant result did not state its CFU count in the published abstract.
Does a higher CFU count mean a better probiotic? No. Dose-response has not been established for most probiotic strains. A trial picks a dose and reports whether it worked; it does not establish that double the dose would work better. A larger number on the box is a marketing variable, not an efficacy one.
How long does L. paracasei take to work? Ninety days is the duration behind the cholesterol result, with measurements taken at day 45 and day 90. Forty-five days is the earliest point the trial looked at, so shorter claims are not supported by that record.
Should I take it with or without food? The cholesterol trial gave the strain in maltodextrin capsules daily and the abstract does not specify timing relative to meals. Any instruction to take it on an empty stomach comes from the product, not from this trial.
Is L. paracasei the same as L. casei? No longer, and the naming matters when you search for trials. The species was reclassified: what older papers call Lactobacillus paracasei is now Lacticaseibacillus paracasei. Searching either name will miss half the literature, and a product label may use either.
Facts checked: 10 November 2026 · VitalRoundup Editorial Team · Reviewed by VitalRoundup Editorial Team (sources and label data re-checked) · How we review
Related notes
Sources
- Safety and Effects of Lactobacillus paracasei TISTR 2593 Supplementation on Improving Cholesterol Metabolism. Thai Clinical Trials Registry TCTR 20220917002. https://pubmed.ncbi.nlm.nih.gov/36771367/
- Evaluating the safety and efficacy of Lacticaseibacillus paracasei TISTR 2593. https://pubmed.ncbi.nlm.nih.gov/39125454/
- Probiotics and strain specificity in lipid outcomes — pooled effect estimates with confidence intervals, collected in our trial effect size reference.
