Author: NP Bate, Ed.S.•Last reviewed: 2026-09-18

Medical information reviewed by NP Bate, Ed.S. • Last reviewed: 2026-08-09 • Not a substitute for professional care.

Treatments We're Watching

This page is our watchlist. These are treatments with a plausible mechanism and some early signals, but not enough evidence to recommend them yet. Most of these are not FDA-approved for the condition listed, and where a device has been authorized we say so and describe what that authorization covers. We update this page as the research moves.

Bottom line up front: mechanism + early data = worth watching. Reproducible, large trials = worth recommending. Every treatment below is in the first bucket.

1. Saffron (Crocus sativus), Macular Health / Early AMD

Watching, the strongest candidate on this page.

The world's most expensive spice, and the one supplement on this page I'd actually spend money on. Its active compounds are crocin and crocetin, antioxidant carotenoids.

Mechanism (plausible)

Crocin and crocetin are potent antioxidants and anti-inflammatory agents. Preclinical work shows they protect retinal pigment epithelium (RPE) cells from oxidative stress, the same stress implicated in AMD. Some evidence also suggests anti-angiogenic activity (relevant to wet AMD).

Evidence

  • RCT, N=25, 20 mg/day, 3 months (crossover). Early AMD. Saffron improved retinal flicker sensitivity vs. placebo. Falsini B, et al. Invest Ophthalmol Vis Sci 2010;51(12):6118-6124. doi: 10.1167/iovs.09-4995
  • Open-label extension, N=29, 20 mg/day, ~14 months. Visual acuity improved ~2 Snellen lines; retinal sensitivity improved 0.3 log units; benefits sustained. Piccardi M, et al. Evid Based Complement Alternat Med 2012;2012:429124. doi: 10.1155/2012/429124
  • RCT, BMJ Open Ophthalmology 2024. Saffron modestly improved multifocal ERG responses in AMD patients, including those already taking AREDS supplements. Broadhead GK, et al. BMJ Open Ophthalmol 2024;9(1):e001399. doi: 10.1136/bmjophth-2023-001399

Caveats

  • Every trial is small (25–93 participants) and short (3–14 months)
  • Endpoints are functional (ERG, acuity)—no trial has shown saffron slows structural progression (drusen, atrophy)
  • Some studies involve researchers with ties to saffron suppliers
  • Doses studied: 20–30 mg/day (≈8–12 threads); retail saffron varies wildly in crocin content—supplement quality is unregulated
  • Cost is significant; crocin-standardized extracts are the only way to approximate study doses

Bottom Line

The only treatment on this page with multiple small RCTs behind it. Genuinely promising, but "multiple small RCTs" is not "proven." If you take AREDS2, saffron may be additive, but don't stop AREDS2 for it.

2. Ginkgo biloba, Macular Health / AMD

Wait-and-see, plausible mechanism, insufficient evidence.

If you've heard of one "brain herb," it's ginkgo. The extract (EGb 761) genuinely increases blood flow to the retina — that part checks out. What's missing is proof that it changes outcomes.

Mechanism (plausible)

Increases retinal and choroidal blood flow, antagonizes platelet-activating factor (inflammation), and has antioxidant activity. Vascular factors and oxidative damage are both implicated in AMD, so the mechanism is biologically coherent.

Evidence

  • Cochrane systematic review (2013): Only 2 randomized trials (119 people total, 6 months, EGb 761 at 60–240 mg/day). Both reported some positive vision effects, but results could not be pooled and adverse effects/QoL were not reported. Conclusion: insufficient evidence. Evans JR. Cochrane Database Syst Rev 2013. doi: 10.1002/14651858.CD001775.pub2

Caveats

  • Two small trials, 20+ years old, no replication since
  • Ginkgo inhibits platelet aggregation: bleeding risk with aspirin, warfarin, or before surgery
  • Retail ginkgo is often not standardized to EGb 761

Bottom Line

The mechanism is real; the human data is a whisper. Not enough to justify the bleeding risk for most people. Watch.

3. Astaxanthin, Macular Health / General Eye Health

Watching, promising mechanism, thin human data.

Salmon, krill, microalgae — this is the pigment that gives them their color. What makes it interesting to eye researchers is that it crosses the blood-retinal barrier, which almost no other carotenoid does.

Mechanism (plausible)

One of the most potent natural antioxidants; anti-inflammatory; improves retinal blood flow in small studies. The ability to reach the retina directly sets it apart from most other carotenoids.

Evidence

  • Review of clinical applications (2020): Several small clinical trials suggest potential benefit in AMD, eye fatigue, and accommodation, but most are small, short, and use surrogate outcomes (blood flow, ERG, symptom scores). Giannaccare G, et al. Mar Drugs 2020;18(5):239. doi: 10.3390/md18050239

Caveats

  • No trial has tested astaxanthin against an AMD progression endpoint (drusen growth, conversion to late AMD)
  • Most positive studies are industry-funded or small single-center
  • No established dose for AMD; common retail doses (4–12 mg/day) are not study-backed for this indication

Bottom Line

A real molecule with a real mechanism. But "reaches the retina" is not "treats AMD." Watching.

4. Bilberry (Vaccinium myrtillus), Night Vision / Eye Health

Skeptical — the famous claim failed rigorous testing.

European blueberries. The whole industry rests on a WWII story about RAF pilots eating bilberry jam to sharpen their night vision, great folklore, bad evidence. The supposed actives are anthocyanins (delphinidins, cyanidins).

Mechanism (plausible in theory)

Anthocyanins are antioxidants; some animal work suggests effects on rhodopsin regeneration and retinal blood flow.

Evidence

  • Systematic review of placebo-controlled trials (2004): Bilberry anthocyanosides do not improve night vision in rigorous trials. The famous claim is not supported. Canter PH, Ernst E. Surv Ophthalmol 2004;49(1):38-50. doi: 10.1016/j.survophthal.2003.10.006
  • No meaningful human data for AMD outcomes.

Caveats

  • The one thing bilberry is famous for (night vision) failed systematic review
  • No AMD-specific trials of consequence
  • Supplement quality varies; anthocyanin content unregulated

Bottom Line

Popular, not promising. Included because it's the most-bought "eye health" herb, and the evidence says the marquee claim doesn't hold. If you enjoy bilberries as food, eat them. Don't buy the pills for vision.

5. Manuka Honey Eye Drops, Dry Eye

Watching, cheap and low-risk, but small studies.

Yes, sterile Manuka honey in an eye drop bottle (Optimel is the main one). It sounds like a gimmick. The mechanism, antibacterial, anti-inflammatory, osmotic, actually fits evaporative dry eye, and it's cheap enough to be worth a conversation with your optometrist.

Mechanism (plausible)

Manuka honey has antibacterial (methylglyoxal), anti-inflammatory, and osmotic activity on the ocular surface. For evaporative dry eye (the 86% type), the antibacterial angle targets lid flora implicated in meibomian gland dysfunction.

Evidence

  • Prospective controlled study (2026): Manuka honey drops associated with greater improvement in post-cataract-surgery dry eye symptoms vs. standard care. García-Bardera J, et al. Front Ophthalmol 2026. doi: 10.3389/fopht.2026.1812914
  • Contact-lens-related dry eye (2017): Optimel Manuka+ well tolerated; dry eye symptoms improved. Wong D, et al. Cont Lens Anterior Eye 2017.
  • Review (2023): Significant improvement in corneal staining and meibum quality reported after Manuka therapy. Prinz J, et al. Pharmaceuticals 2023;16(5):762. doi: 10.3390/ph16050762

Caveats

  • Small samples, short follow-up, mostly post-procedure or contact-lens populations
  • Stinging/burning on instillation is common, the #1 reason people quit
  • Not a first-line therapy; warm compresses + lipid tears + omega-3 remain the proven starting point

Bottom Line

Low-risk, cheap, and the mechanism fits evaporative dry eye. But it stings and the trials are small. If the OTC trifecta hasn't worked after 3 months, it's a reasonable conversation with your optometrist, not a self-prescribed first move.

6. Red Light Therapy (Photobiomodulation), Macular Health & Floaters

Watching. One FDA-authorized in-office device, and pooled trials that find no significant benefit.

Two very different things travel under this name. One is an in-office, prescription-only device with a real FDA authorization behind it. The other is the red panels, masks, and handheld torches sold online, which have none. The evidence applies mostly to the first, and only partly. For floaters it applies not at all.

Mechanism (plausible)

Red and near-infrared light (roughly 590 to 850 nm) is absorbed by cytochrome c oxidase in the mitochondrial respiratory chain. The proposed effect is more efficient energy production and less oxidative stress, and the retina is one of the most mitochondria-dense tissues in the body. That is a real mechanism, not marketing. It also defines the limit: mitochondria sit inside cells, so light applied from outside the eye has no route to break up collagen strands floating in the vitreous gel.

Evidence

  • FDA De Novo authorization (DEN230083): the Valeda Light Delivery System (590, 660, 850 nm) is authorized to improve visual acuity in patients with best-corrected vision of 20/32 to 20/70 who have dry AMD with at least 3 medium drusen, large drusen, or non-central geographic atrophy, and no neovascular or center-involving disease. It is a prescription device, delivered in the clinic as nine sessions over 3 to 5 weeks, repeated every 4 months. FDA's own summary of the benefit: about one line of visual acuity on the eye chart after roughly two years. FDA decision summary (PDF)
  • LIGHTSITE III, 13 months (2024): 100 participants, 148 eyes, double-masked and sham-controlled. It met its primary endpoint: a 2.4-letter difference between groups at month 13 (P = 0.02), with treated eyes gaining 5.4 letters. Boyer D, et al. Retina 2024;44(3). doi: 10.1097/IAE.0000000000003980
  • LIGHTSITE III, 24 months (2026): the primary endpoint was met again at month 21 (+6.2 letters with treatment, P = 0.0036), progression to geographic atrophy was lower in the treated arm (6.8% vs 24.0%, P = 0.007), and no signs of phototoxicity appeared. But only 4.4% of treated eyes reached a gain of 15 letters or more, the size usually needed for a change patients notice, and the trial was run by the device maker. Jaffe GJ, et al. Retina 2026;46(5). doi: 10.1097/IAE.0000000000004822
  • Independent meta-analysis (2025): pooling 6 randomized studies (360 patients, 477 eyes), the authors found no significant improvement in visual acuity (SMD -0.30, 95% CI -0.85 to 0.26, P = 0.30), drusen volume, geographic atrophy, retinal thickness, or microperimetry, with no overall excess of adverse events. Chen KY, et al. Ophthalmol Ther 2025;14(5). doi: 10.1007/s40123-025-01119-w
  • 670 nm at home (2020): 31 people with intermediate AMD plus 11 healthy older adults used a 670 nm torch for 2 minutes daily for 12 months. Night vision improved in the healthy group, and nothing changed in the AMD group, functionally or structurally. This is the design most home devices copy, and it did not work in the condition most buyers have. Grewal MK, et al. J Clin Med 2020;9(4):1001. doi: 10.3390/jcm9041001
  • Single 3-minute exposures (2021): color contrast thresholds improved for about a week after one 670 nm exposure in adults aged 37 to 70. Small study, one research group, a laboratory measurement rather than vision you would notice, and no independent replication. Shinhmar H, et al. Sci Rep 2021;11:22872. doi: 10.1038/s41598-021-02311-1
  • Consumer devices for myopia, a safety signal: two red-light instruments sold for childhood myopia exceeded recognized exposure limits within about 3 minutes of viewing. Ostrin LA, Schill AW. Ophthalmic Physiol Opt 2024;44(2). doi: 10.1111/opo.13272. A case of retinal damage after repeated low-level red-light treatment has also been published: Liu H, et al. JAMA Ophthalmol 2023;141(7). doi: 10.1001/jamaophthalmol.2023.1548. Repeated low-level red light is not FDA-approved in the United States.

Caveats

  • No evidence for floaters. No trials, no case series, and no mechanism to suggest benefit, because a therapy that acts on mitochondria inside cells cannot reach the vitreous collagen that causes floaters.
  • The studied device is not the device being sold. Home panels and masks differ in wavelength, power, beam size, and dose. The authorization covers one prescription device used in a clinic.
  • Dose has a ceiling. Photobiomodulation follows a biphasic dose response, so past a certain point more light does not mean more benefit. The two myopia instruments above are the documented example of overshoot.
  • The authorization is narrow, and it does not replace standard care. It covers a specific AMD stage and vision range. FDA notes that treatment was not studied beyond 54 treatments per eye, that benefit may not persist after stopping, and that patients should avoid treatment within 30 days of photosensitizing drugs. Injections for wet AMD and AREDS2 for progression risk remain the proven moves.

Bottom Line

This is the rare case where the mechanism is real and the marketing still outruns the data. One in-office device is authorized, and FDA describes its benefit as about one line of vision after two years, while independent pooling of the trials finds no significant gain, and the at-home versions that dominate the internet have no authorization and a documented dose risk. For dry AMD it is a conversation with a retina specialist, not a purchase. For floaters, which is what this site is mostly about, it is nothing.

Watchlist Methodology

  • What gets on the list: a plausible mechanism + at least one human study (any design) or a compelling animal/in-vitro mechanism chain. The bar for promotion off the list is: replicated, larger trials with clinically meaningful endpoints.
  • What gets removed: replicated null results, safety signals, or retractions—we update this page and note the change (same policy as our Citations page).
  • What never goes on the list: anything with no biological plausibility, no human data, and a marketing page instead of a methods section.
All Guides
---