The dry-eye trial is sea buckthorn’s best single study: randomized, double-blind, placebo-controlled, about 100 adults, 2 grams of oil daily for three months from autumn into winter (October to January for most participants). It found a smaller seasonal rise in tear film osmolarity and fewer people hitting the worst score for redness and burning, though each of those two results held in only one of the two analyses. It also found no difference in the other diary symptoms or the other clinical tests, and its own follow-up failed to explain the mechanism. Both halves deserve your attention.
See the data behind this chart
| Outcome | Placebo | Sea buckthorn oil (2 g/day) | All participants | Compliant participants only |
|---|---|---|---|---|
| Reached the top redness score | 36% | 6% | Significant (P=0.04) | Not significant (P=0.11) |
| Reached the top burning score (compliant participants) | 32% | 12% | Not significant (P=0.05) | Significant (P=0.04) |
The design, and why it is the good kind
Participants aged 20 to 75 with dry-eye symptoms were randomized to 2 grams of sea buckthorn oil (1 gram twice daily with meals) or placebo for three months, from autumn into winter, the season that punishes dry eyes hardest. The oil was a blend of berry and seed oil, made by Aromtech, a sea buckthorn oil company and one of the trial’s industry partners. Double-blinding meant neither participants nor measurers knew who got what. It is the best-designed study on the research desk.
That design is why this trial anchors the whole dry-eye claim. Randomization kills selection bias, blinding kills expectation effects, and the autumn-into-winter timing gave the outcome room to actually happen.
The seasonal timing deserves a second look, because it is the quiet strength of the design. Dry eyes worsen in cold, dry air, so a three-month run from autumn into winter is a stress test: osmolarity will rise in almost everyone, and the question becomes who rises less. Run the same trial in a mild summer and both groups might drift nowhere, leaving nothing to measure. The season was the intervention’s sparring partner.
The results, including the fussy parts
Tear film osmolarity rose in both groups as winter bit, but the rise was significantly smaller in the sea buckthorn group. The detail worth printing: the effect was P=0.04 on intention-to-treat analysis (everyone randomized, compliant or not) and P=0.02 per-protocol (only the compliers). ITT is the stricter, more demanding lens, and the effect survived it, if narrowly. One more caveat the paper prints: the osmolarity difference was significant only in a model adjusted for baseline osmolarity, baseline Schirmer score and age. Without those adjustments, the authors report, the groups didn’t differ significantly (data not shown). Osmolarity was also one of four main outcome measures. The trial was sized on the mOSDI symptom questionnaire, and neither it nor tear break-up time nor the Schirmer test differed between groups.
The two analyses answer two different questions. Per-protocol asks what the oil did in people who actually took it. Intention-to-treat asks what happened to everyone assigned to it, including the people who skipped doses, which is the question a real-world buyer is really asking. A result that only survives per-protocol lives or dies on perfect compliance. This one survived both analyses, though only with the covariate adjustment.
Maximum redness score: 6 percent of the oil group hit the top score, versus 36 percent on placebo. That cleared significance counting everyone (P=0.04) but not among compliers (P=0.11). Maximum burning: among compliers, 12 versus 32 percent (5 of 43 against 12 of 38) reached the top score, P=0.04, and this is the one number people quote too confidently. Counting everyone assigned, it came in at P=0.05, which is not significant. Contact-lens wearers, a subgroup, had a smaller share of eye-symptom days on the oil when everyone was counted, 65 ± 29 percent of days against 81 ± 26 percent on placebo (P=0.049), but not among the compliers alone, 65 ± 30 percent against 77 ± 27 percent (P=0.19). The diary question was a yes-or-no about any eye symptom, with no intensity, and the oil group’s lens wearers also wore lenses on more days (45 percent of days against 27) and for more hours (14 a day against 11 at the daily maximum, on average), which makes the comparison harder to read.
The other six diary symptoms, soreness, foreign-body sensation, dryness, grittiness, watery eyes and blurry vision, showed no significant difference. The standard clinical measures, tear break-up time, the Schirmer test, and the symptom questionnaire, also showed no difference between groups. Most sites quoting this trial omit that sentence. We keep it, because it tells you the benefit was real but targeted, worst days and a lab measure, not a cure for every sensation of dry eye. The practical version lives on our dry-eye page.
The numbers in everyday terms
Percentages from trials need translating before they mean anything. Of every 100 people on placebo, 36 had a day where redness hit the worst score. Of every 100 on the oil, 6 did. That is a large relative drop and also a statement about the worst days, not an average day, and not a promise that your eyes will feel 30 points better. The burning numbers come from the compliers only: 12 percent on the oil (5 of 43) against 32 percent on placebo (12 of 38) hit the top score, P = 0.04. Counting everyone assigned, the gap came in at P = 0.05, and the shares for that count sit in a table the copy we can read doesn’t include. Redness carries the mirror-image asterisk: it cleared significance counting everyone and not among compliers. The quiet symptoms and tests are the other half of the picture: on most measures, most days, the groups looked alike.
The follow-up that complicated the story
In 2011 the same team analyzed tear-film fatty acids from the trial cohort, expecting the oil’s fatty acids to show up in the tears and explain the benefit. They did not. Their measurement came back flat: the changes in tear-film fatty-acid proportions did not differ between the oil and placebo groups (branched-chain P = 0.49, saturated P = 0.59, monounsaturated P = 0.53, polyunsaturated P = 0.16), so whatever the oil does, the pathway is still unknown. A mechanism is a question mark, and we print the question mark.
That failure-to-find matters more than it looks. The team had every incentive to confirm their own theory, access to the actual trial samples, and a published prediction riding on the outcome. When they reported the fatty acids had not moved, they spent their own credibility on a null. It is the single best reason to trust the positive results from the same lab: this group prints its question marks.
The replication status, stated plainly
One team ran this trial, and the same team ran the follow-up. No independent group has repeated it in the years since, which cuts both ways and only those two ways. The result has not been refuted: no trial has come back with a different answer in people with dry eye. A 2019 eyelid-spray study from the same group reported lower OSDI scores in the treated eye, and we could read only its abstract. The spray is topical and also contains sodium hyaluronate, so it cannot be credited to the oil alone, and it is a follow-on and not an independent replication. The 2024 skin trial did ask its 40 healthy women about eye dryness, and the oil group’s self-rated dryness didn’t differ significantly from placebo’s, but it gave a quarter of the dose in a skin study, so it doesn’t test this result. It has also not been confirmed by fresh hands, and single-lab findings carry single-lab risk. The grade sits at moderate precisely because of this paragraph, and the mechanisms page shows the same team reporting the pathway they failed to find.
What replication would need
- An independent lab, ideally in a different climate, because one team’s result is a lead until someone else finds it too.
- More participants, since P=0.04 on osmolarity, one of four main outcome measures and the only one that differed, leaves little margin.
- All seasons, to separate the winter-specific effect from a general one.
- Mechanism endpoints that test routes beyond tear fatty acids.
One adjacent question rides along: whether people whose dry eyes come from an autoimmune condition would respond the same way. A small pilot in that population circulates in the literature, but it has not cleared our verification bar, so it appears nowhere in our grading. The safety review explains that this trial, like the others we can read, reports no adverse-event or tolerability result.
Until then: one well-run trial, moderate grade, and the studied 2-gram dose on the study dosages page. Two grams a day with meals works out to four 500 mg softgels, so the count on the label matters more than the brand on it. That count is the first thing our best sea buckthorn softgels roundup checks. The human studies in our registry are listed in the clinical trials catalog.
Frequently asked questions
What did the sea buckthorn dry-eye study find?
In a randomized, double-blind, placebo-controlled trial of about 100 adults, 2 grams of sea buckthorn oil daily for 3 months from autumn into winter blunted the seasonal rise in tear film osmolarity (P=0.04 intention-to-treat, P=0.02 per-protocol, in a model adjusted for baseline values and age). The share of people whose worst redness score reached severe was 6 percent on the oil versus 36 percent on placebo (P=0.04 counting everyone, P=0.11 among those who took the capsules on at least 80% of days). Among those compliers, 12 versus 32 percent reached the worst burning score (P=0.04), and that gap missed significance when everyone was counted (P=0.05). The other diary symptoms, tear break-up time, the Schirmer test and the symptom questionnaire showed no significant difference.
Has the dry-eye trial been replicated?
Not by an independent group. The same research team ran a 2011 follow-up on tear-film fatty acids, which found no difference between groups, meaning the mechanism is still unexplained. Its 2019 eyelid-spray study, which we could read only in abstract, is a topical follow-on and not an independent replication. One well-designed trial from one lab is a strong lead, not a settled fact, which is why we grade dry eyes as moderate evidence rather than strong.
How long until the oil helps dry eyes?
The trial's main measurements came at three months, so three months is the only answer the evidence gives for a result. It also ran from autumn into winter on purpose, when dry eyes worsen, which means the benefit showed up as a smaller seasonal decline rather than an overnight fix. Nobody has measured week two, and any product promising instant relief is ahead of the data.
Illustration.