Evening Primrose

Oenothera biennis L., common evening primrose, belongs to Onagraceae, the evening primrose family, which takes its name from flowers that are partly or fully closed by day and open in the evening (USDA). NCBI Taxonomy files it as taxon 3942, alongside many related Oenothera species that hybridize readily. It is not related to the true primroses of European woodlands (Primula). As its name says, it is a biennial: in the first year it forms a flat rosette of leaves at ground level, and in the second it sends up a stiff, often hairy flowering stem anywhere from about 30 cm to 2.5 m tall, sets seed, and dies (USDA). The stem leaves are alternate and lance-shaped, shallowly toothed and often wavy at the edges.

The flowers are bright yellow, 2–5 cm across, with four petals and a cross-shaped stigma; they are fragrant and each lasts only a day or two, opening in sequence up the spike from June into autumn (USDA). NCCIH notes that they open at sunset and close during the day. Night-flying hawk moths, drawn by the scent and pale colour, are thought to be the main pollinators (USDA), with bees and hummingbirds visiting when flowers stay open into a cloudy morning. Each flower leaves a long, narrow capsule packed with small reddish-brown seeds. NCCIH describes the plant as native to the Americas and now growing throughout Europe and parts of Asia, where it has naturalized along roadsides, railway banks, field edges, and other disturbed ground.

The medicinal product comes almost entirely from the seed. EMA defines evening primrose oil as the fatty oil obtained from the seeds of O. biennis or the closely related O. lamarckiana by extraction or expression, and the European Pharmacopoeia standard it cites requires 65–85% linoleic acid (PubChem CID 5280450), 7–14% gamma-linolenic acid (CID 5280933, C18H30O2), and no more than 0.5% alpha-linolenic acid, with smaller amounts of oleic, palmitic, and stearic acids. Linoleic acid is common in seed oils; GLA is not, which is EPO’s whole selling point. Borage (starflower) and blackcurrant seed oils contain even more GLA and appear alongside EPO in many trials. Because these polyunsaturated oils oxidize on exposure to air, light, heat, and moisture, EMA stresses cool, dark storage. The leaves, roots, and stem juice were used in folk medicine but are not the subject of modern products.

Evening primrose was a North American medicine long before it was a supplement. NCCIH records that Native Americans applied juice from the stem and leaves to the skin for inflammation, bruises, and minor wounds, and took the leaves by mouth for digestive complaints and sore throats. EMA’s assessment report says the plant reached Europe early in the seventeenth century, where it was sold under the name king’s cure-all and became a popular folk remedy in Britain before falling out of use for centuries. Herbals described it as astringent and sedative. EMA credits the chemists Heiduschka and Lüft with finding, in 1919, the unusual fatty acid in the seed oil that they named gamma-linolenic acid, and notes that British researchers began studying its effects in animals in the 1960s.

The modern boom began in the late 1970s and early 1980s. Researchers proposed that people with atopic eczema and some other conditions converted linoleic acid to GLA poorly, so supplying GLA directly might correct a biochemical bottleneck; Hywel Williams (BMJ 2003) describes it as a very plausible mechanism that made EPO a welcome alternative at a time of widespread fear of topical steroids. EMA lists the conditions for which EPO products were then marketed: premenstrual syndrome, alcoholism, pregnancy-related high blood pressure, eczema, raised cholesterol, high blood pressure, scleroderma, multiple sclerosis, rheumatoid arthritis, breast pain, and more. In the UK, Scotia Pharmaceuticals’ Epogam was licensed for eczema and Efamast for breast pain, and both were available on NHS prescription.

The licences did not last. Williams records that a 1989 company meta-analysis supporting Epogam relied on seven small company trials that never appeared in the public domain and excluded a larger independent negative trial, and that a 1995 Department of Health–commissioned analysis of 20 studies, including unpublished company data, was never allowed to be published. In 2002 the UK Medicines Control Agency withdrew the marketing authorisations for Epogam and Efamast after concluding that the data, including new studies and statistical analyses, did not support the standard of efficacy required for eczema or breast pain (EMA). There was no safety concern; EPO stayed on sale as a food supplement without medicinal claims. A large independent trial of high-dose GLA in eczema was negative the following year, and Williams titled his BMJ editorial ‘Time to say goodnight’. The EU herbal committee adopted its traditional-use monograph in 2011 and revised it on 5 June 2018; a 2025 review found no new data requiring change.

Linoleic acid from the diet is converted in the body to GLA by the enzyme delta-6-desaturase, then lengthened to dihomo-gamma-linolenic acid (DGLA), the precursor of series-1 prostaglandins such as prostaglandin E1, and further to arachidonic acid. MSKCC explains EPO’s rationale this way: GLA bypasses the first, rate-limiting step, so supplementing it might help people who are poor at converting linoleic acid. Supplementation does raise blood and tissue GLA and DGLA, and MSKCC notes that in some eczema studies these rises tracked symptom changes. Whether that translates into clinical benefit is a separate question, and for eczema the answer from the best trials has been no. Laboratory and animal studies cited by MSKCC report anti-inflammatory effects from the fatty acids and from minor constituents such as long-chain fatty alcohols, but these are early findings.

The same pathway explains the main theoretical safety concern. MSKCC and EMA describe reduced thromboxane formation and increased prostaglandin E1, which together inhibit platelet aggregation; MSKCC cites a small human study in which several months of GLA from evening primrose oil significantly prolonged bleeding time in 9 of 12 people. NCCIH notes that the idea behind using EPO to start labour is that it might increase prostaglandin production, but trials of oral and vaginal EPO for this purpose have had inconsistent results. MSKCC notes that EPO has no oestrogenic activity of its own.

Eczema (atopic dermatitis) is the use with the most trials and the clearest answer. Bamford and colleagues (Cochrane 2013) included 27 randomized studies with 1,596 participants, 19 of evening primrose oil and 8 of borage oil. Pooling seven trials, EPO did not improve global eczema symptoms rated by participants, and pooling eight, it did not improve symptoms rated by doctors, compared with placebo; most studies were judged at low risk of bias. The authors concluded that oral EPO and borage oil are not effective treatments for eczema and that, because the confidence intervals were narrow enough to exclude a clinically useful difference, further studies would be hard to justify. Williams (BMJ 2003) reached the same conclusion after a 151-person trial of high-dose borage oil, which tested the last argument that GLA only works at large doses. NCCIH agrees that oral EPO has not been shown to help. EMA’s traditional-use monograph for itching in dry skin rests on long-standing use, not on proof of efficacy.

Breast pain (cyclical mastalgia) was EPO’s other licensed indication. Ahmad Adni and colleagues (International Journal of Environmental Research and Public Health 2021) pooled 13 randomized trials with 1,752 women and found that the proportion achieving pain relief with EPO did not differ from placebo or from comparators such as vitamin E, topical anti-inflammatory gels, or danazol; adverse events were not increased either. NCCIH, citing that review, says EPO is probably not more effective than placebo for breast pain. An earlier meta-analysis cited by MSKCC likewise found no advantage over placebo, although a small pilot trial suggested benefit when EPO was combined with vitamin E.

Premenstrual syndrome: Budeiri, Li Wan Po, and Dornan (Controlled Clinical Trials 1996) found only seven placebo-controlled trials, in only five of which randomization was clearly described; scoring was too inconsistent to pool, and the two best-controlled studies showed no benefit. They concluded that EPO is of little value for PMS, while noting that small trials cannot exclude modest effects. Menopausal hot flashes: Farzaneh and colleagues (Archives of Gynecology and Obstetrics 2013) randomized 56 women to 500 mg of EPO twice daily or placebo for six weeks; frequency, severity, and duration of hot flashes improved in both groups, and only severity, plus some quality-of-life measures, improved significantly more with EPO. A single six-week trial of this size cannot establish benefit, and NCCIH says the evidence is insufficient for both PMS and menopause symptoms.

Rheumatoid arthritis is the one area with moderately supportive evidence. Cameron, Gagnier, and Chrubasik (Cochrane 2011) pooled seven studies of GLA from evening primrose, borage, or blackcurrant seed oil and found reduced pain intensity (a mean of about 33 points on a 100-point scale) and improved disability compared with placebo, with more adverse events on GLA (20% versus 3%) that did not reach statistical significance. They rated the evidence moderate and called for larger, better-designed confirmatory trials. NCCIH still considers the evidence insufficient. Typical trial doses were around 6 g of EPO a day, providing roughly 540 mg of GLA (EMA), taken alongside standard treatment rather than instead of it.

Diabetic neuropathy: in a one-year multicentre trial, Keen and colleagues (Diabetes Care 1993) gave 111 people with mild diabetic neuropathy 480 mg of GLA a day or placebo and reported that 13 of 16 nerve-function measures changed significantly more favourably with GLA. MSKCC characterizes the early neuropathy studies as equivocal and notes a 2020 Korean trial in which 320 mg of GLA a day was noninferior to alpha-lipoic acid for pain over 12 weeks, with no placebo arm. These findings have not led to a recommended treatment. Other uses, including dry eye, multiple sclerosis fatigue, and ulcerative colitis, rest on small preliminary studies (MSKCC).

Evening primrose oil is sold mainly as soft gelatin capsules, commonly 500 mg or 1,000 mg, and less often as a bottled liquid; it is also an ingredient in skin creams. EMA notes that one gram of oil contains roughly 80–116 mg of GLA, so labels may state GLA content rather than oil weight. Many products add vitamin E as an antioxidant, and some combine EPO with fish oil, borage oil, or, in menopause products, phytoestrogens; MSKCC warns that these combinations change the safety picture. Because the oil oxidizes readily, store capsules in a cool, dark place and do not use products that smell rancid (EMA).

EU traditional use (EMA/HMPC/753041/2017, adopted 5 June 2018): for symptomatic relief of itching in acute and chronic dry skin conditions, adolescents, adults, and older people take 2–3 g of oil as a single dose, 4–6 g per day, by mouth in solid dosage forms. Use under 12 is not recommended because of lack of adequate data. If symptoms persist beyond eight weeks, see a doctor or qualified practitioner. EMA’s assessment report notes that handbooks claim about three months of use is needed before a full effect, a claim that sits awkwardly with the negative trials; there is little point continuing beyond that without clear benefit.

Doses in clinical trials vary with the condition. EMA’s assessment report summarizes them as roughly 4–8 g per day for adult eczema and 2–4.5 g for children, about 3 g for breast pain, 4 g for menopausal symptoms, 4–6 g for premenstrual syndrome, 4–8 g for diabetic neuropathy, and 6 g (about 540 mg GLA) for rheumatoid arthritis. These are research doses, not endorsements, and for most of these conditions the trials were negative. The Farzaneh hot flash trial used only 1 g per day. Midwifery regimens for cervical ripening vary widely and are not recommended; see the cautions on pregnancy.