Practical buying guide
Most people buy probiotics the way they buy cereal. They look at the front of the box, see a friendly phrase like “digestive health,” and pick one. That approach wastes money. Here is the version that works.
Start with your symptom
Probiotics are not interchangeable. Each strain carries its own body of evidence, and that evidence usually points at one or two specific problems. Begin with what you are trying to fix.
If you are about to start antibiotics, the strain with the strongest record is Saccharomyces boulardii CNCM I-745. Multiple randomized trials and pooled analyses support it for preventing antibiotic-associated diarrhea. If your problem is irritable bowel syndrome, Bifidobacterium longum subsp. infantis 35624 has real weight behind it: five pooled trials, 436 people, and the symptom improvement held up. If you have ulcerative colitis and need to hold remission, Escherichia coli Nissle 1917 was tested head to head against mesalazine in 327 patients and performed about the same.
You pick the strain from the problem. The other direction does not work.
Read the strain name, not the species name
This is the most important thing on the label.
A label that says Lactobacillus rhamnosus tells you almost nothing. Dozens of strains sit inside that species and they behave differently. Lactobacillus rhamnosus GG has its own trial record in childhood infectious diarrhea. Lactobacillus rhamnosus HN001 was studied for eczema and for mood. Same species, different strain, different evidence.
A good label gives the full strain designation: genus, species, subspecies if there is one, and the strain code. Something like Lacticaseibacillus rhamnosus GG (ATCC 53103). If a company will not tell you which strain is inside the capsule, that tells you what you need to know about the company.
Check how much evidence the strain has
Every strain on this site carries an evidence score from 1 to 5. Use it.
A 5 means multiple randomized trials plus pooled analyses. A 4 means several trials, with nobody having combined them yet. A 3 means thin human data: one small trial, or a trial where the strain was mixed with others. A 2 means animal or lab work only. A 1 means no direct human evidence at all.
A 1 or a 2 is not automatically useless. Latilactobacillus sakei 23K scores a 2 and it is a genuinely interesting strain for nasal work. It has not been tested in people yet. You deserve to know that before you pay for it.
Match the dose to the trial
This is where most products fail. A strain can have excellent evidence at one dose and no evidence at a lower one.
The landmark IBS trial on B. infantis 35624 used 1 × 10^8 colony-forming units per day. If a product contains the right strain at a fraction of that amount, the trial result does not transfer. Colony-forming units are the live bacteria count. The number on the bottle should be the count guaranteed through the expiration date, not the count at the time of manufacture. Companies that print only the manufacturing count are telling you less than it appears.
Check that it survives the trip
A probiotic has to get through stomach acid and bile and arrive alive. Some strains are built for this. Bacillus coagulans GBI-30, 6086 and Bacillus subtilis DE111 form spores, which are tough survival shells that resist heat and acid. That is why they appear in shelf-stable products and even baked goods. A crossover trial recovered DE111 vegetative cells from the small intestine, which is direct proof of germination in humans.
Other strains are fragile and need refrigeration. If your product shipped hot and sat in a warehouse, the count on the label may not be the count in the capsule.
Be careful with multi-strain blends
A ten-strain product sounds like more value. Often it is less.
If a trial tested one strain alone and you swallow ten, you cannot assume the trial applies. Blends also tend to split a fixed total dose across every strain, so each one lands below its studied amount. Blends have their place. The preterm infant trials used multi-strain formulas and showed real benefit. Check whether the evidence you are leaning on came from the blend or from a single strain inside it.
Ignore the front of the bottle
In the United States, supplements are not approved before they reach the shelf. The FDA does not verify that a probiotic works, or that it contains what the label claims. Front-of-bottle language is marketing. The strain name and the dose on the back are the parts that carry information.
Look for third-party testing seals: USP, NSF, or ConsumerLab. These do not prove the product works. They prove it contains what it says it contains.
Give it time, and actually track it
Most trials run four to eight weeks. One week tells you nothing.
Write down your symptom score before you start. Keep it simple. For congestion, a zero to ten scale every morning. Then check it at week four. Without a baseline you are guessing from memory, and memory is generous to whatever you spent money on.
If your problem is your nose or your sinuses
I want to be straight with you here, because this is the reason I built this site.
The evidence for probiotics in chronic rhinitis and sinusitis is thin. Most strains on this site were studied for gut outcomes: diarrhea, IBS, colitis. A few carry respiratory data, and most of that concerns infection frequency in children rather than chronic congestion in adults. Lacticaseibacillus casei Shirota has trials on upper respiratory infection risk. Bifidobacterium breve M-16V and Bifidobacterium longum BB536 have been studied in allergic rhinitis and asthma, usually in children and usually inside multi-strain formulas.
The nasal microbiome is a young field. The idea that chronic sinus trouble follows a viral hit that lets opportunistic bacteria move in is plausible, and there is early work on it. It is not established. A product claiming to cure your sinusitis is running ahead of the data.
That does not make experimenting pointless. It means you should treat it as an experiment with honest expectations. It is not a proven treatment.
When to stop and see a doctor
Probiotics are safe for most healthy people. There are exceptions. If you are immunocompromised, have a central venous catheter, have short bowel syndrome, or are critically ill, there are published case reports of bloodstream infection from probiotic organisms. Talk to your physician first.
See a doctor too if your symptoms are one-sided, if you have blood in your mucus, if you have facial pain with fever, or if nothing has moved after a genuine eight-week trial. Chronic congestion has structural causes. A deviated septum or a narrow nasal valve will not respond to any bacteria you swallow. I learned that one the hard way.