Probiotic skincare: pool five randomised trials and none of the three measures was significant

A 2026 systematic review and meta-analysis. It gathered randomised controlled trials comparing oral or topical Lactobacillus-based probiotics or postbiotics against placebo or benzoyl peroxide in patients with acne vulgaris.

Included: five randomised trials, 332 participants.

The results.

• Non-inflammatory lesions: pooled mean difference −1.39 (95% CI −5.10–2.32), P = .46
• Inflammatory lesions: −0.08 (−1.28–1.11), P = .89
• Total lesion counts: −9.07 (−20.71–2.57), P = .13

All three confidence intervals cross zero. As written repeatedly elsewhere on this site, an interval crossing zero means neither 'no effect' nor 'an effect' can be excluded.

But the three values differ a little in character.

• Inflammatory lesions: −0.08 with an interval of −1.28–1.11, which is narrow. It clusters tightly around zero, so a large effect is fairly firmly excluded.
• Total lesion counts: −9.07 with an interval of −20.71–2.57, which is wide. The point estimate itself leans toward a sizeable reduction, but the interval is too wide to settle it. This is the place where a larger sample might have changed things.

The authors conclude that Lactobacillus-based probiotics do not provide significant clinical benefits in reducing inflammatory, non-inflammatory or total acne lesion counts compared with placebo or benzoyl peroxide.

Heterogeneity was moderate to low — the pooled trials did not disagree with each other much, which makes the pooled result correspondingly more trustworthy.

And the authors declare no conflicts of interest. That is the opposite condition from the one that kept deoxycholic acid off the top rung on this site for being entirely manufacturer-funded.

So this item is filed as 'not established'. What is not established is whether Lactobacillus reduces acne lesions.

The comparators in the trials this meta-analysis gathered were placebo or benzoyl peroxide. And the pooled result combines the two.

Why that matters.

The two comparators give opposite meanings to 'no difference'.

• No difference against placebo → the probiotic was the same as something inert
• No difference against benzoyl peroxide → the probiotic was the same as a drug known to work

If the second were the case, that would be a good result. Benzoyl peroxide is among the firmest-evidenced acne treatments on this site. Being indistinguishable from it would mean the probiotic did rather well.

But pooling the two erases the distinction. One combined figure remains, and you cannot tell which comparison it came from.

From the abstract alone we cannot verify how many of the five used which comparator. We did not consult the full text, so we do not put that number here.

The direction is guessable, though. Had most been benzoyl-peroxide-controlled, the authors would not have concluded that it 'does not provide significant clinical benefits' — tying with an effective drug is not the same as providing no benefit. So the placebo-controlled trials most likely drove the result.

Even so, that is our inference, not a verified fact. This site writes inference and verification with different words.

On the LED phototherapy page you will have seen the Cochrane authors write that studies using common acne treatments as comparators are needed. Here the two kinds got mixed.

What this meta-analysis gathered were oral or topical Lactobacillus-based probiotics and postbiotics.

Both routes went into one pooled figure.

This site writes repeatedly that a different route needs its own evidence — separating oral from topical collagen on the collagen page, and the supplement from the cream on the l-carnitine page.

Oral and topical probiotics work by entirely different routes.

• Oral — the proposed path runs through the gut flora and systemic immunity. It is a long way to the skin.
• Topical — it reaches the surface flora directly. But putting live bacteria into a cosmetic and shipping it is difficult, so actual products are often lysates of killed bacteria, or ferment filtrates.

That this page sits in the topical ingredients layer makes the problem larger. The pooled figure above has oral trials mixed into it, while anyone arriving at this page here is most likely weighing something to put on their skin.

So what this page can honestly say is:

• Pooling oral and topical Lactobacillus together, there was no significant benefit on acne lesion counts — established
• What topical alone looks like — not established
• Whether live bacteria and lysates or ferment filtrates are the same — not established
• What about uses other than acne (soothing, barrier, sensitivity) — not established. We verified no randomised trial on that for this page

That last line deserves emphasis. In cosmetic advertising, microbiome claims are made more often about soothing and barrier than about acne. The evidence for those claims is not on this page.