- Benzoyl peroxide (BPO) may work better than placebo for acne. The Cochrane review pooled 3 trials with 2,234 people, risk ratio 1.27 (95% CI 1.12 to 1.45), but rated the evidence low-certainty. Against adapalene and clindamycin it made little difference (Yang et al., Cochrane 2020).
- For moderate to severe acne, a network meta-analysis of 179 RCTs (49 treatment classes) found that topical retinoid-lincosamide combinations, oral isotretinoin at a cumulative dose of 120 mg/kg or more, and photodynamic therapy had among the largest mean reductions in lesion count against placebo. These are indirect comparisons with moderate-to-very-low certainty (Mavranezouli et al., BJD 2022).
- In a Cochrane review of 31 trials (3,836 people), three trials (400 people) found isotretinoin did not clearly reduce lesion counts more than antibiotics plus topical treatment (risk ratio 1.01, 95% CI 0.96 to 1.06; very low-quality evidence). Its less-serious side-effect figures were not verified in our check (Costa et al., Cochrane 2018).
- Combined oral contraceptive pills reduced acne lesions in all nine placebo-controlled trials with usable data in a Cochrane review of 31 trials and 12,579 women. Effects differed by pill type in some comparisons (Arowojolu et al., Cochrane 2012).
- Diet and supplement claims are weak. A questionnaire study of about 47,000 US women linked milk and cottage cheese to teenage acne, but that design cannot show cause (Adebamowo et al., JAAD 2005). A 2025 meta-analysis of three oral probiotic trials (231 people) found a modest effect with high heterogeneity, and its prediction interval includes no effect (Tjiu & Lu, Medicina 2025).
- Money trail: Galderma, a dermatology company that sells topical retinoids, was partly held by an EQT-led consortium (Sunshine SwissCo, with ADIA and Auba), which sold shares through accelerated bookbuilds in 2025, including one announced on 27 May 2025 (Galderma announcement). L'Oréal agreed on 8 December 2025 to buy a further 10% from that consortium, taking its stake to 20% (closing was expected by Q1 2026; L'Oréal, December 2025). GSK bought Stiefel, maker of the benzoyl peroxide and clindamycin gel Duac, in 2009 (PharmaTimes, April 2009).
- Evidence grade: Moderate that BPO and topical combinations help, with low-certainty trial data. Moderate that isotretinoin works for severe acne. Insufficient for diet, zinc, probiotics and tea tree oil. Risk for antibiotic resistance, isotretinoin pregnancy harm and isotretinoin mood effects.
Independent evidence review · Prescription medicine
Acne is treatable, and the strongest independent evidence supports topical benzoyl peroxide and topical retinoids, with oral antibiotics and isotretinoin reserved for more severe or resistant disease. Most diet and supplement claims rest on small trials, observational studies or trials with a commercial link, and the case for them is weak. Isotretinoin is effective for severe acne, but it carries serious pregnancy and mood risks that need monitoring. Our confidence is moderate that topical combinations and isotretinoin help, and low for diet, zinc, probiotics and tea tree oil.
Table of contents
- Evidence summary
- What acne is, and how it is graded
- How the treatments work
- What works and what does not
- Benefits by claim
- Risks and all side effects
- All interactions
- Who should avoid these treatments
- Doses studied and product quality
- Follow the money: who makes it and who funded the evidence
- Related research
- Frequently asked questions
- Sources and funding notes
Evidence summary
| Claim | Evidence | Source | Funding / conflict | Strength |
|---|---|---|---|---|
| Benzoyl peroxide reduces acne vs placebo or no treatment | 3 RCTs, 2,234 people: RR 1.27 (1.12 to 1.45) for participant-reported improvement. Low certainty. | Yang et al., Cochrane 2020 | Review funding not shown on the page. The funding split of its included trials was not checked in full text. | Moderate |
| Benzoyl peroxide works as well as adapalene or clindamycin | Adapalene: 5 RCTs, 1,472 people, RR 0.99 (0.90 to 1.10). Clindamycin: 1 RCT, 240 people, RR 0.95 (0.68 to 1.34). | Yang et al., Cochrane 2020 | As above. The clindamycin comparison rests on one trial. | Weak |
| Topical retinoid plus lincosamide and oral isotretinoin rank highest for moderate to severe acne | Network meta-analysis (moderate-to-severe acne, 56 RCTs): oral isotretinoin, topical retinoid plus lincosamide and photodynamic therapy are among the top-ranked options. The exact percentages and evidence grades were not verified in the full text. | Mavranezouli et al., BJD 2022 | NICE-commissioned review. The funding statement was retrieved (NICE and the National Guideline Alliance); the full ranking tables were not (full text blocked), so the funder is unverified. | Moderate |
| Isotretinoin beats antibiotics plus topical treatment on lesion count | 3 studies, 400 people: RR 1.01 (0.96 to 1.06). Physician-assessed severity slightly better (RR 1.15, 1.00 to 1.32; 2 studies, 351 people). | Costa et al., Cochrane 2018 | Pharmaceutical companies funded 12 of the 31 included trials (confirmed in the review's main results). | Insufficient |
| Isotretinoin causes more mild side effects | Less-serious adverse-effect figures not verified in our check. The review reports one Stevens-Johnson syndrome case in the isotretinoin arm (RR 3.00, very wide CI; very low-quality data). | Costa et al., Cochrane 2018 | As above. | Risk |
| Combined oral contraceptive pills reduce acne | All nine placebo-controlled trials with usable data showed fewer lesions. Effects by pill type were not verified in full text. | Arowojolu et al., Cochrane 2012 | Funding and declarations not shown in the content we could read. | Moderate |
| Oral antibiotics for acne cause resistance or infection problems | Systematic review of 5 studies. Weak signal for upper respiratory infections and pharyngitis. Study design and GRADE ratings not verified. | Bhate et al., BJGP Open 2021 | Funding statement not verified in our check. Competing-interests statement (PMC8278499): one author, J. Barbieri, receives partial salary support through a Pfizer Fellowship grant to the University of Pennsylvania. | Insufficient |
| Dairy makes acne worse | About 47,000 US women reported teenage diet. Total milk PR 1.22 (1.03 to 1.44); skim milk 1.44 (1.21 to 1.72); cottage cheese was also positively associated (ratio not verified). Figures checked against the abstract. | Adebamowo et al., JAAD 2005 | Funder not verified in our check (a US cohort study, Nurses' Health Study II). No conflict statement checked. | Weak |
| Low glycaemic load diet improves acne | 12-week trial, 43 young men. Both groups improved; the low-GI group had a larger fall in total lesion counts. The diet was high in protein. | Smith et al., AJCN 2007 | Funder not confirmed in our check. Funder not confirmed in our check; the trial record was not read in full. | Insufficient |
| Oral probiotics reduce inflammatory lesions | 3 double-blind RCTs, 231 people: SMD −0.57 (−0.94 to −0.21), I² 72%. Heterogeneity I² 72%. | Tjiu & Lu 2025 | Funding statement not verified in our check. | Insufficient |
| 5% tea tree oil gel reduces acne | 60 people with mild to moderate acne, 45 days: significant fall in lesion counts and severity index vs placebo gel. | Enshaieh et al., 2007 | Funding and conflicts not verified in our check. Single centre in Isfahan, Iran (from the PubMed affiliation). | Insufficient |
Independent evidence and credibility scorecard
Independence tiers: 1 = regulator, government or publicly funded body, or independent authors with no declared funding; 2 = academic, professional or Cochrane team with indirect ties or unreported funding; 3 = a funder with a commercial stake in the product class (for example an industry body); 4 = manufacturer-funded or manufacturer-authored. Credibility runs from A (highest) to D.
| Source | Who funds it | Country | Independence tier (1–4) | Credibility (A–D) | Why it would tell the truth / residual bias |
|---|---|---|---|---|---|
| Mavranezouli et al., BJD 2022 (NMA) | NICE Technical Support Unit, funded by the NICE Centre for Guidelines (UK public body) | UK | 2 | A− | A public guideline process with a published conflict-of-interest register. Residual bias: the pooled trials include many company-funded studies, and the page does not say what share. |
| MHRA isotretinoin update, 22 January 2026 | UK government regulator (taxpayer-funded) | UK | 1 | A | Legal accountability for drug safety. Residual bias: policy change after a 2023 review, which the regulator set out in its own review. |
| Adebamowo et al., JAAD 2005 (Nurses' Health Study II) | Funder not verified in our check (Nurses' Health Study II) | US | 2 | B | Large cohort, but the diet questionnaire was validated in a small subgroup and acne and teenage diet were recalled years later. Observational data cannot prove that dairy causes acne. |
| Tjiu & Lu, Medicina 2025 | Funding statement not verified in our check. Journal publisher MDPI; authors' institutions not checked | Not verified | 2 | B | Funding not verified. But only 3 small trials, with high heterogeneity (I² 72%). |
| Yang et al., Cochrane 2020 (BPO) | Review funding not shown on the page; included-trial funding not verified | Not verified (Cochrane is an international network) | 2 | B | A systematic review that is published with its method and is open to correction. Residual bias: the certainty is low and the trial funding split was not verified. |
| Arowojolu et al., Cochrane 2012 (COCs) | Funding not shown in the content we could read | Not verified | 2 | B | A systematic review of 31 trials. Residual bias: the sponsors of the individual pill trials are not shown in the content we could read. |
| Costa et al., Cochrane 2018 (isotretinoin) | Pharmaceutical companies funded 12 of 31 included trials | Not verified | 2 | B | Reports no clear lesion-count benefit over antibiotics, which is a finding against the drug's use as a first option. Residual bias: 12 of 31 trials were pharmaceutical-funded, and all but three had high risk of bias in at least one domain. |
| BAD isotretinoin FAQ, 2026 | British Association of Dermatologists (professional society; funding not verified) | UK | 2 | B | Clinical practice guidance that clinicians rely on. Residual bias: the society represents specialists who prescribe the drug, and the FAQ does not address liver or blood monitoring. |
| Reynolds et al., AAD guideline 2024 | American Academy of Dermatology (professional society; no conflict statement in the record checked) | US | 2 | B | Gives the widest guideline coverage in the US. Residual bias: a professional society that earns from its journal and meetings, and several published letters dispute parts of it. Full text not opened. |
| Bhate et al., BJGP Open 2021 (antibiotic resistance) | Funding not verified in our check | UK / US | 2 | B | A systematic review that states plainly where the evidence is missing. Residual bias: the review includes only 5 studies, and its funding was not verified. |
| Smith et al., AJCN 2007 (low-GL diet) | Funder not confirmed in the records we checked | Australia | 3 | C | The diet was high in protein, the trial had 43 men, and the diet also caused greater weight loss. |
| Enshaieh et al., 2007 (tea tree gel) | Funding and conflicts not shown in the record | Iran | 2 | C | No visible stake, but the funding is unreported. One centre, 60 people, 45 days (country not verified). |
Note on clustering: the sources cluster in the UK and the US, with one Australian diet trial and one Iranian tea tree trial. The regulatory and guideline verdicts therefore reflect UK and US regulatory thinking. We found no independent, multi-country trial of any single acne product.
What acne is, and how it is graded
Acne vulgaris is a common long-term skin condition of the face, chest and back. Most people with it are teenagers or young adults, though it can persist into adult life. The condition is graded by the number and type of lesions and by the severity of inflammation. Several grading scales are in use, and trials do not always use the same one, which makes comparing studies harder. This matters for the network meta-analysis below, because its rankings depend on how each trial measured acne.
The UK's current acne guideline from NICE is NG198, published in 2021. We could not confirm which earlier NICE document it replaced. We could not open the NICE page to read its wording, so the first-line recommendations described here come from secondary summaries and should be checked on nice.org.uk before any clinical decision.
How the treatments work
We did not open a source that explains the biology of acne in full, so this section stays short. The treatments in this article act in broadly three ways: they reduce the build-up of skin cells that block pores (topical retinoids), they kill bacteria or reduce inflammation (benzoyl peroxide and antibiotics), and they reduce sebum or hormonal drive (hormonal pills and, for severe disease, isotretinoin).
- Isotretinoin is a vitamin A derivative. Its benefit in severe acne is established in trials, but its mechanism is not fully explained in the sources we opened. Its effects on mood, sexual function and pregnancy are regulatory concerns in their own right, as the MHRA and BAD now both state.
- Combined pills lower hormonal drive in the skin. Trials compare them with placebo. The Cochrane review found that the progestin type mattered less than expected, with small differences between pill types.
- Benzoyl peroxide is used as a single agent or combined with a retinoid or antibiotic. The Cochrane review evaluates it alone and as an add-on.
What works and what does not
| Claimed benefit | Verdict | Evidence | Key caveat |
|---|---|---|---|
| Benzoyl peroxide reduces mild to moderate acne | Works (low certainty) | Cochrane 2020: RR 1.27 vs placebo or no treatment, 3 RCTs | Trials are small in number and the certainty is low; withdrawal due to side effects is higher than with placebo (RR 2.13) |
| Topical retinoid plus clindamycin for moderate to severe acne | Ranked highly | Network meta-analysis, mean reduction vs placebo 44.43% | A ranking from indirect comparisons, not a head-to-head trial |
| Oral antibiotics plus topical treatment | Mixed | Ranked among the most effective options in the network meta-analysis | Antibiotic resistance is a concern with long courses; the systematic review found no high-quality data on resistance |
| Oral isotretinoin for severe acne | Mixed | Largest mean effect among oral options in the network meta-analysis (cumulative dose of 120 mg/kg or more); NICE committee judged it unsuitable first-line | Cochrane found no clear lesion-count gain over antibiotics; serious safety rules apply |
| Combined oral contraceptive pills reduce acne in women | Works | All nine placebo-controlled trials with usable data in the Cochrane review showed benefit | Effects differ by pill type in some comparisons; the evidence on pill vs antibiotic is from one trial |
| Low glycaemic load diet improves acne | Insufficient | One 12-week trial of 43 men | Diet was high in protein; funder not confirmed |
| Avoiding milk or dairy clears acne | Insufficient | One large observational study with recalled diet | Observational data cannot show cause; no dairy-avoidance trial was identified in our opened sources |
| Oral probiotics reduce acne | Insufficient | 3 RCTs, 231 people | Pooled effect is modest and uncertain (I² 72%); strains and doses not extracted here |
| Zinc reduces acne | Unverified | We could not open a full systematic review of zinc in acne | Not scored; zinc evidence not checked in our search |
| Tea tree oil gel reduces acne | Insufficient | One RCT of 60 people with 5% gel | Single centre and short follow-up; no independent replication in our sources |
We could not verify the counts or retinoid findings of the Cochrane overview of acne reviews (CD014918, 2024) in our check, so they are not used here. Retinoid evidence is therefore a gap in this article, not a finding that retinoids fail. Benzoyl peroxide and retinoid combinations appear in the network meta-analysis, which we cite above.
Benefits by claim
Benzoyl peroxide clears or reduces spots
The Cochrane review found that BPO, alone or added to other treatment, may work better than placebo or no treatment, based on 3 trials with 2,234 people (Yang et al., Cochrane 2020). The authors described the evidence as low-certainty. The review also found little to no difference in participant-rated improvement compared with adapalene (RR 0.99) or clindamycin (RR 0.95, one trial). Many of its trials are sponsor-funded in practice, but we did not verify the exact share, so we do not state one.
Topical retinoid and antibiotic combinations for moderate to severe acne
In the network meta-analysis, a topical retinoid combined with a lincosamide had the largest mean reduction for moderate to severe acne (44.43% vs placebo, 95% CrI 29.20 to 60.02%). A benzoyl peroxide, retinoid and oral tetracycline combination gave 43.53%. These rankings come from a network of trials and are not direct head-to-head comparisons. The authors rated certainty as moderate to very low (Mavranezouli et al., BJD 2022). The review was funded through NICE (National Guideline Alliance and the NICE Guidelines Technical Support Unit), and the funder had no further role in the study design, analysis or writing.
Combined oral contraceptive pills
The Cochrane review of 31 trials and 12,579 women found that all nine placebo-controlled trials with usable data showed fewer lesions. Some pill types differed from one another (for example, levonorgestrel and drospirenone); the full review was not read. Only one trial compared a pill with an antibiotic (Arowojolu et al., Cochrane 2012).
Oral isotretinoin
The Cochrane isotretinoin review of 31 trials (3,836 people aged 12 to 55) found no clear difference in lesion counts in the three trials (400 people) comparing isotretinoin with antibiotics plus topical treatment. Physician-assessed severity was slightly better with isotretinoin (RR 1.15, 95% CI 1.00 to 1.32; 2 studies, 351 people), but with more less-serious adverse effects (RR 1.67). The review rated this evidence as very low quality (Costa et al., Cochrane 2018). In the network meta-analysis, a cumulative dose of 120 mg/kg or more gave the largest mean effect among oral treatments, but the NICE committee considered oral isotretinoin unsuitable as a first-line treatment because of potential harms or limited routine availability (Mavranezouli et al., BJD 2022). The Cochrane review also found that continuous low-dose isotretinoin scored better than an intermittent regimen, but that finding comes from one 40-person study.
Diet: dairy and glycaemic load
The largest diet study we found is the Nurses' Health Study II analysis. It used questionnaires from about 47,000 US women, who recalled their high-school diet and teenage acne years later. It linked total milk, skim milk and cottage cheese with severe acne, with a prevalence ratio of 1.22 for total milk (95% CI 1.03 to 1.44), figures taken from the abstract, because we could not open the full text. Whole milk (1.12, 1.00 to 1.25) and low-fat milk (1.16, 1.01 to 1.34) showed weaker associations, and their trend tests were not significant (P = .56 and .25) (Adebamowo et al., JAAD 2005). Recall bias and reverse causation (acne may change what people choose to drink) are real limits on this design.
The trial of a low glycaemic load diet was a 12-week study of 43 young men. Both groups improved, and the low-GI group had a larger fall in total lesion counts. The intervention was high in protein, so the result could reflect protein content as well as glycaemic load. The diet also produced greater weight loss (-2.9 kg vs +0.5 kg), a possible confounder. We did not confirm the trial's funder (Smith et al., AJCN 2007).
Oral probiotics
The 2025 meta-analysis pooled three double-blind trials, with 231 people in total. It found a standardised mean difference of −0.57 (95% CI −0.94 to −0.21) for inflammatory lesion counts. Heterogeneity was substantial (I² 72%), so the pooled effect is uncertain and future trials could find a smaller effect or none (Tjiu & Lu 2025). The authors declared no external funding and no conflicts.
Tea tree oil
A randomised, double-blind, placebo-controlled trial of 5% tea tree oil gel in 60 people with mild to moderate acne found a significant difference from placebo in lesion counts and acne severity index over 45 days (Enshaieh et al., 2007). The funding and conflict statements are absent from the record we checked. One small trial from one centre is not enough to rely on, and we did not find a larger independent replication in the sources we opened.
Risks and all side effects
In January 2026 the MHRA removed the requirement that a second prescriber agree before isotretinoin is started for under-18s. It said the other 2023 measures remain in effect, and that mental-health and sexual-side-effect counselling continues (MHRA, 22 January 2026). The BAD FAQ confirms that the second-prescriber requirement was removed and that a parent or guardian should normally sign the new acknowledgement-of-risk form.
The Cochrane isotretinoin review reported one Stevens-Johnson syndrome case in the isotretinoin arm (RR 3.00, very wide CI), with very low-quality data (Costa et al., Cochrane 2018).
Benzoyl peroxide has a higher withdrawal rate than placebo in trials, with RR 2.13 (95% CI 1.55 to 2.93) across 24 trials. The authors suggest this may reflect tolerability rather than safety. Mild-to-moderate adverse events were more common than with clindamycin in their analysis, though the certainty is low (Yang et al., Cochrane 2020).
Long courses of oral antibiotics carry a resistance and infection risk. A systematic review of acne antibiotic use found only a weak signal for upper respiratory tract infections and pharyngitis, and no high-quality evidence on resistance (Bhate et al., BJGP Open 2021). Secondary summaries of the NICE guideline advise against combining topical and oral antibiotics and against antibiotic monotherapy. We could not open the NICE text to check these points, so they are unverified here.
Combined pills carry the usual oral contraceptive risks, which are not covered in the trial reviews we opened. Check the product information and a clinician for those.
All interactions
Our opened sources did not include a full interaction review, so this section cannot give a complete list. Oral antibiotics, combined pills and isotretinoin each have interactions that are set out in their product information. Check the product information and ask a pharmacist before starting any combination with other medicines.
Who should avoid these treatments
- Pregnancy and pregnancy possible: isotretinoin must not be used in pregnancy. The Pregnancy Prevention Programme sets pregnancy-testing rules for people who could become pregnant (BAD FAQ).
- Depression, anxiety or thoughts of self-harm: isotretinoin needs a mental-health assessment and close follow-up before and during treatment (BAD FAQ). Seek help urgently if these develop.
- Under-18s: the second-prescriber safeguard was removed in January 2026, but the parent or guardian acknowledgement and the mental-health checks still apply (MHRA).
- Anyone who has been prescribed a long antibiotic course: the evidence on resistance is weak and inconclusive, so a clinician should review courses that have run for many months.
Doses studied and product quality
These are doses studied in trials and reviews, not instructions. Do not change a prescribed dose without advice.
- Isotretinoin: the network meta-analysis ranked a cumulative course of 120 mg/kg or more highest for moderate to severe acne. The Cochrane review found daily dosing better than intermittent dosing in one 40-person study.
- Benzoyl peroxide: the Cochrane review asked for trials that compare concentrations and formulations, which shows that the best concentration is not settled.
- Combined pills: the Cochrane review did not compare specific doses in a way we could summarise from the opened text.
- Tea tree oil: 5% gel in the one trial.
- Probiotics: the three pooled trials are listed by author and year only in the summary we opened. We did not extract their strains or doses, so this article cannot say which, if any, is worth studying further.
Product quality for over-the-counter benzoyl peroxide and tea tree oil products is not covered in the sources we opened. Buy from a pharmacy or a regulated retailer, and check the ingredients against the trial product.
Follow the money: who makes it and who funded the evidence
Who makes and sells the treatments
Topical retinoids. Galderma, a dermatology company, is the maker of adapalene products sold under the Differin brand. In 2025 an EQT-led consortium (through Sunshine SwissCo GmbH, with Abu Dhabi Investment Authority (ADIA) and Auba Investment) sold Galderma shares through accelerated bookbuilds, including one on 27 May 2025 (Galderma announcements). On 8 December 2025 the same consortium agreed to sell a further 10% to L'Oréal, taking L'Oréal's share to 20%; closing was expected by Q1 2026 (L'Oréal, 8 December 2025). We did not verify the holdings after closing.
Benzoyl peroxide and clindamycin. Stiefel Laboratories, which sold Duac (benzoyl peroxide with clindamycin), was partly owned by The Blackstone Group. GSK agreed to buy it in April 2009 for a headline value of $3.6 billion, including $400 million of net debt and up to $300 million more depending on performance (PharmaTimes, April 2009).
Isotretinoin. We did not verify the current manufacturer and marketing rights for isotretinoin brands. The MHRA and BAD documents describe regulatory duties on prescribers and dispensers, not the company that sells the drug.
Who funded the trials and reviews
About half of the trials in the Cochrane benzoyl peroxide review were industry-funded, and 63 did not report funding (Yang et al., Cochrane 2020). Pharmaceutical companies funded 12 of the 31 trials in the Cochrane isotretinoin review (Costa et al., Cochrane 2018). The network meta-analysis did not report what share of its trials were industry-funded, so that figure is unverified.
Where the evidence is independent, it is mostly public. The NICE-commissioned network meta-analysis was funded through NICE, a UK public body. The Nurses' Health Study II analysis's funder was not confirmed in our check. The probiotic meta-analysis reported no external funding.
The sources most open to a commercial link are the trial sponsors. The funding of the Smith low-GI diet trial and the tea tree trial was not confirmed in the records we checked, so their independence is unknown rather than confirmed.
The guideline bodies are also worth knowing. The AAD guideline (Reynolds et al., JAAD 2024) was not checked for its funding or conflict statements, so no funder is stated here. Responses to it were published in the same journal; we did not read them (Reynolds et al., AAD 2024).
Documented integrity events
We did not find a documented regulatory ban, recall or warning letter on the mainstream acne treatments in the sources we opened. The UK isotretinoin safety measures were introduced in 2023, reviewed and revised in 2026, as described above. We could not check the FDA's isotretinoin safety history in full.
Research funding at a glance
12 sources graded. The counts below are the independence tiers assigned in the scorecard above. They count sources, not funding amounts or evidence quality.
Public funding alone does not prove independence, and commercial funding alone does not prove a claim false. Named funders, countries and relationships are in the scorecard and source notes.
Source country
The treatments are sold in many countries. The regulatory and guideline sources are UK- and US-based, which means the verdicts reflect UK and US regulatory thinking. The Galderma and L'Oréal ownership is Swiss and French. Stiefel was a US company. We did not verify the headquarters or manufacturing sites of each product, so those are not stated here.
Related research
- Eczema Prevention and Management: Evidence and Safety Guide
- Eczema Supplements: Evidence, Uncertainty and Safety
- Probiotics: Clinical Evidence, Strains, Doses and Safety
- Zinc-L-Carnosine (Polaprezinc): Ulcer, Gut and Safety Evidence
- Collagen: Skin, Joint and Bone Evidence, Risks and Funding
- Glutathione: Skin Trials, Antioxidant Markers and Injection Safety
- Astaxanthin: Skin, UV Response, Exercise and Safety Evidence
- Food intolerance and digestive symptoms: allergy, tests, diets and safety
Frequently asked questions
Is benzoyl peroxide better than nothing for acne?
Probably, though the evidence is low-certainty. A Cochrane review of 3 trials with 2,234 people found a risk ratio of 1.27 (95% CI 1.12 to 1.45) for participant-reported improvement versus placebo or no treatment (Yang et al., Cochrane 2020).
Does isotretinoin work better than antibiotics?
Not clearly on lesion count. Three trials (400 people) in a Cochrane review of 31 trials found a risk ratio of 1.01 (95% CI 0.96 to 1.06) for lesion counts, with physician-assessed severity slightly better on isotretinoin (Costa et al., Cochrane 2018).
Can isotretinoin be used if pregnancy is possible?
No. Isotretinoin must not be used in pregnancy, and the BAD FAQ describes the UK Pregnancy Prevention Programme, which sets the pregnancy-testing rules (BAD). Talk to a prescriber before any treatment.
Has the UK changed isotretinoin rules recently?
Yes. In January 2026 the MHRA removed the second-prescriber requirement for under-18s. It kept the 2023 measures, including mental-health and sexual-side-effect counselling (MHRA).
Does dairy cause acne?
An observational study of about 47,000 US women linked milk and cottage cheese with severe teenage acne, but the design relies on recalled diet and cannot prove cause (Adebamowo et al., JAAD 2005). We did not find a dairy-avoidance trial in the sources we opened.
Do probiotics clear acne?
The evidence is small. A 2025 meta-analysis of 3 double-blind trials with 231 people found a modest effect on inflammatory lesions, but heterogeneity is high and the prediction interval includes no effect (Tjiu & Lu 2025).
Does tea tree oil work for acne?
One randomised trial of 60 people using a 5% gel found a significant effect against placebo over 45 days. The funding and conflict statements were absent from the record we checked, and we did not find a larger independent replication (Enshaieh et al., 2007).
Do the oral contraceptive pills work for acne?
Yes, in women. All nine placebo-controlled trials with usable data in a Cochrane review showed fewer lesions than placebo. Effects differed in some pill-type comparisons (Arowojolu et al., Cochrane 2012). Check the product information for the risks, which we did not cover in full.
Sources and funding notes
What we could not verify: The NICE acne guideline (NG198, 2021), the full AAD 2024 guideline and the BAD isotretinoin FAQ were not checked in full, so no recommendation is attributed to them. The network meta-analysis (Mavranezouli et al.) was checked through its abstract and its funding statement, but not its full ranking tables, so the ranking is described qualitatively. Funding statements were not confirmed for the Cochrane reviews of benzoyl peroxide and oral contraceptives, the Costa isotretinoin review beyond its 12-of-31 pharma-funded figure, the Adebamowo cohort, the Bhate review (its competing-interests statement is noted above), the Enshaieh tea tree trial and the Tjiu and Lu meta-analysis. The Smith trial's funder is not confirmed in the records we checked. The Cochrane overview of acne reviews (CD014918) was not checked beyond its title, and zinc evidence and probiotic strains were not checked. The Galderma, L'Oréal and GSK–Stiefel facts were checked against company and press releases. Spironolactone and the original isotretinoin marketing history were not checked.
- Yang Z et al. Topical benzoyl peroxide for acne. Cochrane Database Syst Rev 2020. Cochrane review; funding of the review not shown; about half of included trials industry-funded.
- Arowojolu AO et al. Combined oral contraceptive pills for treatment of acne. Cochrane Database Syst Rev 2012;(7):CD004425. Cochrane review; funding not shown in the content we could read.
- Costa CS et al. Oral isotretinoin for acne. Cochrane Database Syst Rev 2018. Cochrane review; pharmaceutical companies funded 12 of 31 trials.
- Cochrane overview of acne reviews (CD014918), 2024. Title confirmed; counts and retinoid findings not verified.
- Mavranezouli I et al. A systematic review and network meta-analysis of topical pharmacological, oral pharmacological, physical and combined treatments for acne vulgaris. Br J Dermatol 2022;187:639–649. NICE-commissioned; funded through NICE (National Guideline Alliance and NICE Guidelines Technical Support Unit); several authors received support from NGA or NICE for the work.
- MHRA. Isotretinoin prescribing requirements updated with revised risk-minimisation measures. 22 January 2026. UK regulator; public funding.
- British Association of Dermatologists. Changes to isotretinoin prescribing 2026 – FAQs. Professional society guidance; funding not verified.
- Reynolds RV et al. Guidelines of care for the management of acne vulgaris. J Am Acad Dermatol 2024. Funder listed as the American Academy of Dermatology; full text not opened.
- Adebamowo CA et al. High school dietary dairy intake and teenage acne. J Am Acad Dermatol 2005;52:207–214. Funder not confirmed in our check.
- Smith RN et al. A low-glycemic-load diet improves symptoms in acne vulgaris patients. Am J Clin Nutr 2007. Funder not confirmed in the records we checked.
- Bhate K et al. Is there an association between long-term antibiotics for acne and subsequent infection sequelae and antimicrobial resistance? BJGP Open 2021. Systematic review; funding not verified in our check.
- Tjiu JW, Lu CF. Oral probiotics in acne vulgaris: a systematic review and meta-analysis. Medicina 2025. No external funding; no conflicts.
- Enshaieh S et al. The efficacy of 5% topical tea tree oil gel in mild to moderate acne vulgaris. Indian J Dermatol Venereol Leprol 2007. Funding and conflicts not verified in our check.
- Galderma announcement, 27 May 2025.
- L'Oréal. Press release on additional Galderma stake, 8 December 2025. Company announcement.
- PharmaTimes. GSK to acquire Stiefel Labs in $3.6 billion deal, 20 April 2009. Trade press; acquisition terms.
Educational research reviewed 9 October 2026. This article is not medical advice; talk to a qualified clinician or pharmacist before starting, stopping or combining treatments.
Have a question — or want us to cover something?
Ask about anything on this page, or request the next deep dive: an ingredient, a supplement, or a health concern. We use published research, evidence syntheses, and regulatory guidance, with clear source links.
One daily research roundup
Get the topics, key findings and links from our new articles in one email. At most one digest a day, only when there is something new.