Chinese Herbs for Dry Eyes: What 33 Trials Actually Show

Dry eyes are one of the signature discomforts of modern life. Screens, air conditioning, contact lenses, and age conspire against the tear film, and for the many people whose eyes burn, itch, or tire by late afternoon, the default advice is depressingly thin: use artificial tears, blink more, take breaks.

Artificial tears help — they lubricate. But they do not pretend to change the underlying condition, and anyone who has gone through bottle after bottle knows the relief is temporary. So when a research team published, on October 6, one of the largest attempts ever to rank Chinese herbal formulas for dry eye disease by their trial evidence, it deserved a careful look. Not a credulous one. A careful one.

What the Researchers Did

A team from the Hong Kong Polytechnic University and the Centre for Eye and Vision Research searched five databases — PubMed, Embase, the Cochrane Library, CINAHL, and China's CNKI — for randomized controlled trials of Chinese herbal medicine for dry eye disease, covering everything up to mid-2025. They found 33. The protocol was registered in advance with PROSPERO (CRD420251072475), which is the methodological equivalent of showing your work.

The cleverest design choice was how they grouped the treatments. Individual Chinese herbal formulas are endless — decoctions are customized, patent medicines multiply — so the team classified each trial's intervention into one of five pattern-based classes reflecting the classical diagnostic framework: Yin–Kidney (YK), Liver–Spleen Regulating (LSR), Qi–Yin Support (QYS), Yin-Moistening (YM), and Blood-Nourishing/Heat-Clearing (BHC), plus a "modern" category. This is not arbitrary taxonomic tidying. It mirrors how these formulas are actually used: chosen by pattern, not by brand name. Testing the classes rather than 33 different product names is a fairer test of the underlying therapeutic idea.

The outcomes were three standard eye measurements: the Schirmer I test (how many millimeters of tear are produced in five minutes), tear film break-up time (how many seconds the tear film stays intact — a measure of tear stability), and corneal fluorescein staining (how much surface damage shows under dye). The method was a Bayesian network meta-analysis: the tournament approach that can rank treatments even when not every pair was directly compared, plus a meta-regression asking whether sicker patients benefit more, and a formal grading of how confident the evidence is.

What They Found

No single class dominated. The winners varied by outcome, which is the first thing to notice:

• Tear production (Schirmer): Yin–Kidney and Qi–Yin Support formulas showed the largest gains over standard care — roughly 2.6 mm and 5.3 mm more than standard care, respectively.
• Tear stability (break-up time): Yin–Kidney and Yin-Moistening formulas improved this by about 2.0 and 2.5 seconds.
• Surface damage (fluorescein staining): Yin–Kidney formulas showed the largest reduction, followed by Liver–Spleen Regulating formulas.

The team also graded each comparison's certainty with the CINeMA framework and assessed bias in every trial with the standard Cochrane tool. And their meta-regression found that patients who started out worse — lower tear production, more surface damage — tended to improve more. Baseline tear stability did not show this pattern.

Three Reasons to Read Carefully

1. The most eye-catching number is also the shakiest.

Qi–Yin Support formulas reportedly gained 5.3 mm of tear production — the largest single estimate in the paper. But its 95% credible interval stretches from about 0.6 all the way to 10.1. That is an enormous band of uncertainty: the true effect could be transformative or barely noticeable, and the data cannot say which. When a ranking declares a treatment "probably best" while its interval spans an order of magnitude, the honest reading is that the treatment is promising and the evidence is thin. Rankings are probabilities, not verdicts.

2. The gains are real but modest.

About 2.5 mm more tear flow on the Schirmer test. About two extra seconds of tear stability. These are genuine improvements over standard care in randomized trials — but they are incremental, and the paper does not establish what change patients actually notice. In dry eye research, small measured improvements do not always translate into "my eyes feel better by dinner." Note also that the Yin–Kidney interval for surface damage just touches zero — the boundary of no effect. A careful reader does not round that up.

3. Herbal trials are hard to blind, and the literature is largely Chinese.

A pill versus a placebo can be blinded cleanly. A bitter herbal decoction cannot be disguised easily. The paper notes that many trials used the cleanest available architecture — herbs plus artificial tears against placebo plus artificial tears, isolating the extra effect of the herbs — but expectation effects cannot be fully separated from pharmacology, especially on the more subjective outcomes. And five databases sounds global, but with CNKI in the mix, this evidence base is overwhelmingly Chinese. That is context, not dismissal: the same caveat applies to many Western-dominated trial literatures. We examined a similar dynamic in the chronic pharyngitis trials.

The Pattern Worth Watching

The meta-regression finding — worse baseline, bigger improvement — is the part of this paper most worth sitting with, because it has two completely different explanations, and both might be partly true.

The flattering explanation: treatments genuinely help the sickest most, because there is simply more room to improve. A patient whose tear film is already near normal cannot gain much; a patient with severe deficiency can. This is what a real effect looks like.

The unflattering explanation: regression to the mean. In any repeated measurement, the worst initial readings tend to improve on retest purely by statistical gravity, even without treatment. Randomized trials control for this at the trial level — the control group regresses too — but in a meta-regression across trials, baseline severity can still confound the comparison.

Which is it? The honest answer is that a meta-regression cannot fully separate them, and the authors do not claim otherwise. What makes this finding interesting rather than suspicious is its specificity: the moderation appeared for tear production and surface damage but not for tear stability. A pure statistical artifact would be less choosy. That asymmetry is the kind of detail that makes a follow-up trial worth designing.

There is also a classical resonance, stated without therapeutic claims. In the traditional framework, dry, burning eyes are classically associated with depleted yin of the liver and kidney — and the Yin–Kidney class, which topped two of the three outcomes, contains formulas like Lycium–Rehmannia pills, historically prescribed exactly for this image of dryness. Whether the old diagnostic map is pointing at something real, or the classification simply grouped the most-studied formulas, is an open question. The trial data were not designed to test pattern theory — but the alignment is worth noticing.

The Bottom Line

This is a serious synthesis: 33 randomized trials, a registered protocol, five databases, bias assessment, and a formal grading of evidence certainty. It finds that Chinese herbal formulas, used alongside standard care, improve standard dry eye measurements by modest amounts — and that different classical pattern classes appear to help different aspects of the condition, roughly in line with what the traditional framework would predict.

That last observation cuts both ways. If the pattern-based ranking is a coincidence of a small, heterogeneous literature, larger trials will wash it out. If it holds, it would be evidence that the classical diagnostic categories are tracking something physiologically real. The most useful thing this meta-analysis does is not crown a winner — it tells the field exactly which classes, on which outcomes, are worth putting to a large, well-designed, long-follow-up test.

If you are interested in how the classical Chinese medical tradition thinks about health more broadly — its ideas of balance, prevention, and the body's patterns — our book Huangdi Neijing: A Modern Introduction offers a beginner-friendly guide to the foundational text of TCM.

This article is for educational purposes only and is not medical advice. If you have persistent dry eye symptoms, please consult a qualified healthcare professional.

References

• Hu W, Jiang Y, Wong MS, Zhou L. Effects of Traditional Chinese Medicine for Prevention and Management of Dry Eye Disease: A Systematic Review, Network Meta-Analysis and Meta-Regression of Randomized Controlled Trials. Nutrients, 2026, 10(4), 76. https://www.mdpi.com/2411-5150/10/4/76
• PROSPERO registration CRD420251072475. https://www.crd.york.ac.uk/PROSPERO/view/CRD420251072475