← Back

Eyes and vision

Everything in the library studied for it, medical and holistic together. Within each kind, what was shown to help comes first, strongest evidence first. What was tested and not shown to help stays on the list.

Strong several reviews agree · Moderate one review or consistent trials · Limited one or a few small studies · Untested no study in people, listed with what it rests on

Medication3 Peptides and SARMs1 Supplements5 Practices2 Daily life2 Lab markers5

Medication

Ezetimibe and fenofibrateModerate

Fenofibrate: diabetic eye disease

Shown to slow diabetic retinopathy in people with early eye changes, 22.7% against 29.2% over 4 years in the largest eye trial, without changing vision.

Compared with placebo [12][13]; placebo or observation [14]

The numbers and sources

Effect: Referable retinopathy or maculopathy, or treatment for it, occurred in 22.7% on fenofibrate 145 mg and 29.2% on placebo (HR 0.73, 95% CI 0.58 to 0.91), with no effect on visual function, quality of life or acuity [12]. Pooling three trials, retinopathy progression fell (RR 0.77, 95% CI 0.68 to 0.88), 45 fewer per 1,000, and no trial found a benefit for vision [13]. An earlier Cochrane review found little to no difference in a mixed group (RR 0.86, 95% CI 0.60 to 1.25) but slower progression in people with retinopathy at the start (RR 0.21, 95% CI 0.06 to 0.71) [14]. In the main heart trial, laser treatment was needed by 3.6% on fenofibrate and 5.2% on placebo [8].
Population: 1,151 adults with diabetes and early retinal changes in Scotland, median 4 years [12]; 3,756 people in 3 trials [13]; 15,313 people with type 2 diabetes in 2 trials [14]; 9,795 adults [8].
Sources: [12], [13], [14], [8]
Caveats: Mostly type 2 diabetes. The earlier Cochrane review predates the largest eye trial [14].

InsulinModerate

Eye, kidney and nerve complications in type 1 diabetes

Shown to cut new eye disease by about 76%, kidney protein leakage by 39% to 54% and nerve damage by 60% with intensive insulin treatment in type 1 diabetes.

The numbers and sources

Effect: In people with type 1 diabetes and no eye disease, intensive insulin treatment lowered the risk of developing retinopathy by 76% (95% CI 62 to 85) against conventional treatment. In people with mild retinopathy it slowed progression by 54% (95% CI 39 to 66). Across both groups it lowered microalbuminuria by 39% (95% CI 21 to 52), albuminuria by 54% (95% CI 19 to 74) and clinical neuropathy by 60% (95% CI 38 to 74) [1]. A Cochrane review of 12 trials found retinopathy developed in 6.2% on intensive control against 23.2% on conventional control (RR 0.27, 95% CI 0.18 to 0.42), nephropathy in 16.3% against 28.4% (RR 0.56, 95% CI 0.46 to 0.68) and neuropathy in 4.9% against 13.9% (RR 0.35, 95% CI 0.23 to 0.53) [2].
Population: 1,441 people with type 1 diabetes, assigned to an insulin pump or three or more daily injections, or to one or two daily injections [1]; 2,230 patients in 12 trials followed for 1 to 6.5 years [2].
Sources: [1], [2]
Caveats: The trials compared more intensive with less intensive insulin treatment. Neither arm went without insulin. The largest trial dominates the review [2]. Once complications were present, the effect was weaker for eye disease and uncertain for kidney disease [2].

SulfonylureasLimited

Eye, kidney and other small-vessel complications

One trial found 25% fewer small-vessel complications over 10 years with intensive sulfonylurea or insulin treatment than with diet in newly diagnosed type 2 diabetes.

The numbers and sources

Effect: Against diet, intensive treatment with a sulfonylurea or insulin lowered any diabetes-related endpoint by 12% (95% CI 1 to 21) and microvascular endpoints by 25% (95% CI 7 to 40), including the need for retinal laser treatment. Results did not differ between chlorpropamide, glibenclamide and insulin. Diabetes-related death (10% lower, 95% CI -11 to 27) and death from any cause (6% lower, 95% CI -10 to 20) did not differ significantly [4].
Population: 3,867 newly diagnosed patients with type 2 diabetes, median age 54, followed over 10 years [4].
Sources: [4]
Caveats: The trial pooled sulfonylureas with insulin in its intensive group, and used chlorpropamide, glibenclamide and glipizide. It did not find fewer large-vessel events [4].
Why limited: This rests on one trial, of 3,867 people. Strength counts independent studies that agree, so one trial is limited at any size.

Peptides and SARMs

KPVGrey marketUntested

Wound healing (eye and mouth)

Not studied in people. KPV eye drops sped healing of scraped corneas in rabbits, and a gel containing KPV aided healing of mouth ulcers in rats.

No study in people. This rests on animal studies, lab studies.

The numbers and sources

Effect: After the corneal surface was scraped off in both eyes, rabbits given KPV eye drops had smaller remaining defects than those given saline. At 60 hours, 8 of 8 corneas treated with KPV or with a nitric oxide donor had fully healed over, against none given placebo. Blocking nitric oxide production inhibited the effect of KPV, and in cultured rabbit corneal cells KPV raised cell viability [15]. In rats with chemotherapy-induced mouth ulcers, a hydrogel carrying KPV and EGCG, an antibacterial compound, improved food intake, body weight recovery and tissue repair, and lowered IL-1beta and TNF-alpha [16].
Population: Rabbits, and rabbit corneal cells in the lab [15]. Rats with chemotherapy-induced oral mucositis [16].
Sources: [15], [16]
Caveats: The gel also contained EGCG, which has its own antibacterial activity, so its effect cannot be assigned to KPV alone [16]. Scraped corneas and mouth ulcers in animals do not show what happens to wounds in people.

Supplements

AstaxanthinLimitedStudied in children

Eye strain

One trial found 4 mg a day lowered eye strain scores 20% more than placebo in children with heavy screen time. Several authors worked for the maker.

The numbers and sources

Effect: In children with daily screen time of 4 hours or more, 4 mg a day for 84 days lowered computer vision syndrome questionnaire scores by 4.00 points against 1.72 on placebo, a 20% between-group difference. Visual fatigue scores were 27% lower than placebo. Visual acuity, near point of accommodation and the Schirmer tear test did not differ between groups [9].
Population: 64 children aged 10 to 14 with mild to moderate eye strain symptoms [9].
Sources: [9]
Caveats: One trial. Several authors were employed by Fuji Chemical Industries, which makes AstaReal astaxanthin, or its subsidiaries, and others by a contract research organization Fuji engaged [9].

Hyaluronic acid (oral)Limited

Dry eyes

One small trial found oral hyaluronic acid added to hyaluronic acid eye drops improved corneal staining and symptom scores more than the drops alone over 3 months.

The numbers and sources

Effect: In 54 adults with dry eye using hyaluronic acid eye drops, those also given 240 mg of oral hyaluronic acid a day had better corneal staining at 1 month (between-group p < 0.001) and better corneal staining and OSDI symptom scores at 3 months (p < 0.001 and p = 0.046). Symptom scores fell from 61.8 to 42.3 on the combination. Tear break-up time did not differ between groups [10].
Population: 54 adults with dry eye disease in Korea, 24 given oral hyaluronic acid and 30 drops alone [10].
Sources: [10]
Caveats: A pilot trial with no placebo capsule, so the control group knew it was not taking a supplement. Everyone also used hyaluronic acid drops [10]. It was publicly funded and the authors declared no conflicts [10].

SeleniumLimited

Mild Graves' eye disease

One trial found better quality of life and slower eye disease progression over 6 months in people with mild Graves' eye disease.

The numbers and sources

Effect: After 6 months, 100 µg twice a day improved quality of life, reduced eye involvement and slowed progression against placebo. The Clinical Activity Score fell more on selenium. Results held at 12 months [8].
Population: 159 people with mild Graves' orbitopathy [8].
Sources: [8]
Caveats: One trial. The abstract gives P values without effect sizes. Funded by the University of Pisa and the Italian ministry for education and research [8].

Vitamin AModerateStudied in pregnancy

Pregnancy outcomes

Not shown to lower maternal or newborn death or preterm birth in pregnant women. Night blindness and anemia in mothers were lower.

The numbers and sources

Effect: Vitamin A alone did not change maternal death (RR 0.88, 95% CI 0.65 to 1.20), perinatal death (RR 1.01, 95% CI 0.95 to 1.07) or preterm birth (RR 0.98, 95% CI 0.94 to 1.01). Maternal night blindness (RR 0.79, 95% CI 0.64 to 0.98) and maternal anemia (RR 0.64, 95% CI 0.43 to 0.94) were lower [2].
Population: 19 trials with over 310,000 women, mostly in Nepal, Ghana and Bangladesh [2].
Sources: [2]
Caveats: Baseline vitamin A status differed between populations. The reduction in maternal infection (RR 0.45, 95% CI 0.20 to 0.99) came from low quality evidence [2].

Practices

Red light therapyModerateStudied in children

Childhood nearsightedness

Shown to slow myopia progression in children by 0.68 diopters per 6 months with repeated low-level red light, at low certainty and with very high variation.

The numbers and sources

Effect: Across 8 studies pooled from 13 (1,857 children and adolescents), repeated low-level red light slowed refractive change by 0.68 D per 6 months (95% CI 0.38 to 0.97, I² 97.7%) and axial lengthening by 0.35 mm per 6 months (95% CI -0.51 to -0.19) compared with control [27].
Population: 1,857 children and adolescents, in 8 RCTs, 3 non-randomized trials and 2 cohort studies [27].
Sources: [27]
Caveats: The authors rate the evidence low certainty and call for 2-year trials [27]. This is a desktop laser instrument viewed with the eye, a different thing from a skin panel. See the retinal injury block under Risks.

Red light therapyModerate

Vision in dry macular degeneration

Not shown to improve vision in dry age-related macular degeneration at 12 months in a Cochrane review of 2 small trials.

The numbers and sources

Effect: Corrected visual acuity at 12 months did not differ meaningfully between photobiomodulation and sham or no treatment (mean difference 0.02 logMAR, 95% CI -0.02 to 0.05, 2 RCTs, 90 eyes). Conversion to the exudative form did not differ (risk ratio 0.97, 95% CI 0.17 to 5.44) [34].
Population: 90 participants with non-exudative age-related macular degeneration in 2 single-center RCTs [34].
Sources: [34]
Caveats: Low to very low certainty. One trial was unmasked, and investigators in the other reported conflicts of interest [34]. Searches ran to May 2020.

Daily life

SunlightStrongStudied in children

Myopia in children

Shown to cut new myopia in children by about 16% to 46% with more time outdoors, in three meta-analyses including several school trials.

Compared with control groups in trials and children spending less time outdoors in cohorts [5]; control groups in trials [6][7]; children in control schools who kept their usual activity [8]; children in control schools [9]

The numbers and sources

Effect: Outdoor time cut incident myopia in trials (RR 0.536, 95% CI 0.338 to 0.850) and cohorts (RR 0.574), and reduced myopic shift by 0.30 diopters over 3 years. In children already myopic, time outdoors was not related to progression [5]. Across 5 randomized trials, new myopia fell (RR 0.76, 95% CI 0.67 to 0.87), with no heterogeneity [6]. Across 7 randomized trials, myopia incidence fell (RR 0.84, 95% CI 0.78 to 0.91) and axial length grew 0.09 mm less [7]. An added 40-minute outdoor class each school day cut 3-year myopia incidence from 39.5% to 30.4% [8]. A recess-outdoors program slowed myopic shift in both myopic and non-myopic children over 1 year [9].
Population: 25 studies, 23 in children [5]. 3,014 children aged 6 to 12 in 5 trials, 9 to 36 months [6]. 9,437 children and adolescents in 7 trials [7]. 1,903 children aged about 6.6 in Guangzhou, China [8]. 693 first-graders in Taiwan [9].
Sources: [5], [6], [7], [8], [9]
Caveats: Most trials were in East Asia, where myopia is common. Outdoor time is daylight plus distance viewing, so sunlight is not the only possible cause. One trial found an effect at light levels as low as shade under trees [9]. The reviews disagree on children already myopic [5][9].

Screen timeModerate

Blue-light filtering glasses

Not shown to ease eye strain from computer use against ordinary lenses in short trials, with effects on sleep uncertain, in a Cochrane review of 17 small trials.

Compared with non-blue-light filtering lenses [8]

The numbers and sources

Effect: Visual fatigue scores may not differ at under 1 week (mean difference 9.76, 95% CI -33.95 to 53.47, 120 participants, low certainty). Visual acuity probably did not differ (MD 0.00 logMAR). Of six trials of sleep, three found better sleep scores and three found no difference (very low certainty). Reported side effects included headache, discomfort and lower mood [8].
Population: Adults in 17 randomized trials of 5 to 156 participants, followed under 1 day to 5 weeks [8].
Sources: [8]
Caveats: No trial had a low risk of bias in every domain, and no meta-analysis was possible [8].

Lab markers

Hemoglobin A1c (HbA1c)Strong

Eye, kidney and nerve complications of diabetes

Shown in trials: intensive glucose control that lowered HbA1c about 0.9 points cut kidney events 20% and eye events 13% in type 2 diabetes.

The numbers and sources

Effect: In an individual-participant meta-analysis of 4 trials in type 2 diabetes, more intensive control left HbA1c 0.90 percentage points lower (95% CI 0.58 to 1.22) and reduced kidney events (HR 0.80, 95% CI 0.72 to 0.88) and eye events (HR 0.87, 95% CI 0.76 to 1.00), with no reduction in nerve events (HR 0.98, 95% CI 0.87 to 1.09) [11]. In type 1 diabetes, intensive therapy reduced the development of retinopathy by 76% (95% CI 62 to 85) and clinical neuropathy by 60% (95% CI 38 to 74) [12]. In an observational analysis of people with type 2 diabetes, each 1% lower updated mean HbA1c went with 37% fewer microvascular complications (95% CI 33 to 41), with no threshold [13]. In pooled data from nine studies, diabetes-specific retinopathy was rare below 6.0% and its threshold fell at 6.3 to 6.7% [14].
Population: 27,049 adults with type 2 diabetes in 4 trials, median follow-up 5.0 years [11]. 1,441 people with type 1 diabetes followed a mean 6.5 years [12]. 3,642 people with type 2 diabetes [13]. 28,010 people with HbA1c and retinal photographs [14].
Sources: [11], [12], [13], [14]
Caveats: The trials tested treatment strategies, not HbA1c itself. Intensive therapy raised severe hypoglycemia two to threefold in the type 1 diabetes trial [12].

Fasting glucoseModerate

Diabetic eye disease

Linked to diabetic eye disease from about 6.4 to 6.8 mmol/L (about 115 to 123 mg/dL) upward, in pooled data on 41,411 people.

The numbers and sources

Effect: In pooled data from nine studies, the prevalence of diabetes-specific retinopathy was low at fasting glucose below 6.0 mmol/L and rose above that level. The threshold fell at 6.4 to 6.8 mmol/L by one method and 6.6 mmol/L by another. The authors suggested the diagnostic level for fasting glucose could be lowered from 7.0 to 6.5 mmol/L [15].
Population: 44,623 people aged 20 to 79 with gradable retinal photographs in nine studies from five countries, 41,411 of them with fasting glucose [15].
Sources: [15]
Caveats: Cross-sectional. WHO states that its diabetes criteria mark a group with more premature death and more microvascular and cardiovascular complications, while noting the limits of the data behind them [5].

Vitamin A (retinol) levelModerateStudied in children

Death in young children where deficiency is common

Shown in trials to cut deaths 12% in children aged 6 months to 5 years given vitamin A in countries where deficiency is common.

The numbers and sources

Effect: In a Cochrane review, vitamin A supplementation reduced deaths from any cause (RR 0.88, 95% CI 0.83 to 0.93, 19 trials, 1,202,382 children, high-certainty evidence) and deaths from diarrhea (RR 0.88, 0.79 to 0.98). It reduced night blindness (RR 0.32, 0.21 to 0.50), Bitot's spots (RR 0.42, 0.33 to 0.53) and vitamin A deficiency itself (RR 0.71, 0.65 to 0.78, 4 studies, 2,262 children). Deaths from measles and from respiratory disease did not differ [8].
Population: 47 studies with about 1,223,856 children in 19 countries, 30 of the studies in Asia. The mean age was about 33 months [8].
Sources: [8]
Caveats: The trials gave vitamin A to children in the community whatever their own serum retinol, so they show what supplementation did in populations with widespread deficiency. They did not test treating a child according to a measured level, and they do not describe adults or populations where deficiency is rare [8].

C-peptideLimited

Eye disease and low blood sugar in type 1 diabetes

Limited evidence: in one type 1 diabetes trial cohort, higher stimulated C-peptide went with less retinopathy and hypoglycemia, and fasting C-peptide showed weaker links.

The numbers and sources

Effect: In the intensive treatment group of the Diabetes Control and Complications Trial, a 50% higher stimulated C-peptide at entry, such as 0.15 against 0.10 nmol/L, went with a 25% lower risk of sustained retinopathy, an 8.2% lower risk of hypoglycemia, a lower insulin dose and an HbA1c 0.07 percentage points lower. Outcomes changed in a nearly linear way as C-peptide fell to unmeasurable levels, with no threshold. Fasting C-peptide had weaker associations with outcomes than stimulated C-peptide [9].
Population: Participants in the Diabetes Control and Complications Trial with up to 5 years of type 1 diabetes at entry, followed for 7 years [9].
Sources: [9]
Caveats: The main findings are for C-peptide measured after a stimulus, a different quantity from the fasting value on this page. They are unadjusted, come from one trial cohort, and concern people with type 1 diabetes only. The abstract gives no participant count [9].

FructosamineLimited

Eye and kidney complications

One US cohort linked fructosamine above the 95th percentile to 50% more new chronic kidney disease, and higher levels to more retinopathy.

The numbers and sources

Effect: Fructosamine and glycated albumin were strongly associated with retinopathy (P below 0.0001 for trend). The adjusted hazard ratio for chronic kidney disease was 1.50 (95% CI 1.22 to 1.85) for fructosamine above the 95th percentile and 1.48 (1.20 to 1.83) for glycated albumin, compared with people without diabetes below the 75th percentile. Prediction of chronic kidney disease was nearly as strong as with HbA1c, with C statistics of 0.717, 0.717 and 0.726 [5].
Population: 11,348 adults without diabetes and 958 with diagnosed diabetes, followed for two decades [5].
Sources: [5]
Caveats: Observational, from one cohort. A laboratory guideline states that few published studies have convincingly shown a relationship between glycated protein levels and the chronic complications of diabetes [2].