Diabetic retinopathy is the eye complication of long-standing diabetes, and it is one of the leading causes of vision loss in working-age adults. This site already covers inherited retinal degeneration and age-related macular degeneration. Diabetic retinopathy is a separate literature with a separate mechanism, and it deserves its own page rather than a sentence inside someone else’s. The diabetes side of the site is split the same way, including type 1 diabetes and the beta cell.
The short version is that TUDCA has been studied in three diabetic animal models and two cell models with broadly consistent results, that the most useful finding in the whole set is about timing rather than dose, and that the only human data are observational measurements of bile acids in blood rather than trials of TUDCA as a treatment. Nothing on this page is medical advice, and diabetic eye disease is monitored by an eye specialist for good reasons.
Key Takeaways
- In diabetic mice, TUDCA started one week after diabetes was induced preserved spatial frequency and contrast sensitivity at every timepoint measured. Treatment started three weeks after induction preserved less[1].
- The same study found reduced a-wave, b-wave and oscillatory potential amplitudes in diabetic mice, along with delayed a-wave and oscillatory potential timing. Early TUDCA improved all three amplitude deficits[1].
- In a mouse model combining diabetes with retinal TNF overexpression, TUDCA reduced vascular leakage and restored visual function, and the effect disappeared when the TGR5 receptor was silenced in human retinal endothelial cells[2].
- In human retinal microvascular endothelial cells under high glucose, TUDCA lowered nitric oxide and reduced ICAM-1, nitric oxide synthase, NF-kappaB p65 and VEGF[3].
- Two observational studies in people with type 2 diabetes measured serum bile acids and found them linked to retinopathy status. Neither gave anyone TUDCA[4][5].
- There is no clinical trial of TUDCA for diabetic retinopathy in humans.
Why a Chemical Chaperone Is a Plausible Candidate Here
Diabetic retinopathy is classified as a microvascular disease, and screening looks for vascular signs: microaneurysms, haemorrhages, new vessel growth. But the retina is nervous tissue, and a body of work argues that neural damage begins before the vascular signs are visible. That matters for a compound like TUDCA, whose mechanism is chaperone activity in the endoplasmic reticulum rather than anything specifically vascular.
Prolonged high glucose stresses the protein folding machinery of the cell. In cultured rat retinal neural cells exposed to elevated glucose for an extended period, TUDCA markedly decreased cell death[6]. The mechanism the authors identified is worth noting because it is not the caspase pathway that dominates most TUDCA writing: TUDCA partially prevented apoptosis-inducing factor from leaving the mitochondria and accumulating in the nucleus, and it reduced two markers of oxidative damage, protein carbonyl groups and reactive oxygen species[6].
So the case is that hyperglycaemia damages retinal neurons through stress pathways TUDCA is known to blunt, and that this happens early enough to be worth intervening in. That is a mechanistic case, not a clinical one.
The Timing Result
The single most practically relevant experiment in this literature is a 2021 study in mice made diabetic with streptozotocin[1]. The design included something most supplement research leaves out: two different treatment start times.
Control and diabetic mice were treated with vehicle or TUDCA beginning either one week or three weeks after diabetes was induced. Visual function was tested bimonthly using an optomotor response, and retinal function was tested monthly with scotopic electroretinography[1].
Diabetic mice showed significantly reduced spatial frequency and contrast sensitivity thresholds compared with controls. Diabetic mice given TUDCA early showed preservation at all timepoints[1]. On the electroretinogram, a-wave, b-wave and oscillatory potential 2 amplitudes fell in diabetic mice, and a-wave and oscillatory potential timing was delayed. Early TUDCA improved the a-wave, b-wave and oscillatory potential deficits[1].
Late treatment showed reduced preservation compared with early treatment[1]. That is the sentence to sit with. A two-week difference in start time, in a mouse, changed how much function was saved. Human diabetic retinopathy is usually identified at a screening appointment, by which point the process has been running for years. The animal result that looks best is the one furthest from how the human situation actually presents.
The Vascular Side and the TGR5 Question
A 2023 study used transgenic mice that overexpress TNF in the vasculature, then made them diabetic, which produces both inflammation and hyperglycaemia in the same animal[2]. Compared with wild-type mice, these animals had decreased visual function that tracked with a decrease in protein kinase C alpha in rod bipolar cells, and increased vascular permeability that was worse again when diabetes was added. Subcutaneous TUDCA over four weeks alleviated these changes[2].
The receptor work is the interesting part. Inflammation and endoplasmic reticulum stress in the retina rose alongside increased expression of TGR5, the bile acid receptor, and TUDCA brought the TGR5 expression back down[2]. In human retinal endothelial cells challenged with TNF plus high glucose, TGR5 expression more than doubled, leukocyte migration across the cell layer tripled, and permeability rose. TUDCA reversed all of that[2].
Then the authors silenced TGR5 and repeated the experiment. Cells without TGR5, given either TUDCA or a TGR5 agonist, failed to reverse the damage[2]. That is a proper mechanistic test rather than a correlation, and it identifies a specific receptor this effect depends on.
A 2016 study reached a similar place from a different direction. In human retinal microvascular endothelial cells under high glucose, TUDCA at 5, 25 and 125 micromolar lowered nitric oxide, and in diabetic rats given 250 or 500 mg/kg per day it lowered serum nitric oxide[3]. Both immunohistochemistry and western blot showed reduced ICAM-1, nitric oxide synthase, NF-kappaB p65 and VEGF[3]. VEGF is the target of the injectable drugs that currently treat proliferative diabetic retinopathy, which makes that particular marker worth flagging, and also worth not overreading.
What the Parent Compound Adds
TUDCA is ursodeoxycholic acid conjugated with taurine, and the two compounds are not interchangeable. UDCA has its own diabetic retinopathy study: in streptozotocin-diabetic mice, UDCA at 15 and 30 mg/kg reversed blood-retinal barrier breakdown and restored expression of the tight junction proteins claudin-1 and claudin-19[7]. It also reversed thinning of both the inner and outer nuclear layers, reduced the microglial marker Iba1, and reduced TNF alpha, IL-1 beta, IL-6, ICAM-1, inducible nitric oxide synthase and VEGF[7].
That is a coherent picture across two related molecules, which is mild evidence that the effect is real rather than a fluke of one lab. It is not evidence that either works in people.
The Human Data, Which Is Not What It Looks Like
Two studies have measured bile acids in the blood of people with type 2 diabetes and related them to retinopathy. Neither administered anything.
The first compared 82 healthy controls, 58 people with type 2 diabetes and retinopathy, and 60 with type 2 diabetes and no retinopathy, with a further 40 patients used for validation[4]. Of fifteen bile acids and eleven unsaturated fatty acids measured, three separated the two diabetic groups: taurolithocholic acid, TUDCA and arachidonic acid. Together they predicted retinopathy with an area under the curve of 0.785, and 0.918 in the validation cohort, and all three remained independent predictors after adjustment for confounders[4].
Note what that does and does not say. It says circulating TUDCA differs between diabetics with and without retinopathy strongly enough to help classify them. It does not say the difference is protective, and the abstract does not report which group had more. A biomarker that discriminates can be a cause, a consequence, or a bystander.
The second study looked at 336 people with type 2 diabetes: 229 who had gone more than ten years without retinopathy and 107 with proliferative retinopathy[5]. Total bilirubin and total bile acid were significantly higher in the group without retinopathy, while triglycerides, cholesterol and LDL cholesterol were higher in the proliferative group. As bile acid level rose across groups, the proportion with proliferative retinopathy fell, and logistic regression classified bilirubin and bile acid as protective factors while systolic blood pressure, LDL cholesterol and urinary albumin excretion rate were risk factors[5].
This is the most encouraging human number available, and it is about total bile acid, not TUDCA. People with better metabolic control differ from people with worse control in dozens of measurable ways at once. Observational adjustment reduces that problem without eliminating it.
What This Evidence Can and Cannot Support
It can support the statement that TUDCA protects retinal structure and function in diabetic rodents, that the effect has an identified receptor dependency, and that starting earlier worked better than starting later in the one study that tested both.
It cannot support any claim about human vision. There is no randomised trial, no open-label trial, and no case series of TUDCA for diabetic retinopathy. Doses in the rodent work run to 250 and 500 mg/kg per day[3], and typical human supplement doses are nowhere near a body-weight-scaled equivalent. Two of the studies used injection rather than oral administration[2][3], which bypasses the absorption question entirely.
The practical point is unglamorous. Glycaemic control, blood pressure control and regular dilated eye screening are the interventions with human outcome data behind them, and proliferative disease has established treatments that work. A supplement with mouse data is not a substitute for an appointment, and anyone with diabetes who notices a change in vision needs an eye examination rather than a purchase.
Frequently Asked Questions
Can TUDCA prevent diabetic retinopathy?
Unknown in humans. In diabetic mice, TUDCA started one week after diabetes onset preserved visual and retinal function at all measured timepoints, and starting at three weeks worked less well[1]. No human has been given TUDCA for this purpose in a published trial.
Is there human evidence linking bile acids to diabetic retinopathy?
Yes, but it is observational. One study of 336 people with type 2 diabetes found higher total bile acid in those without retinopathy and classified bile acid as a protective factor in regression analysis[5]. Another found TUDCA among three metabolites that distinguished diabetics with and without retinopathy[4]. Measuring a molecule in blood is not the same as giving it as a drug.
Does TUDCA lower VEGF, the target of eye injections?
In cells and rats, yes. TUDCA reduced VEGF expression in high-glucose human retinal microvascular endothelial cells and in diabetic rat retinas[3], and UDCA did the same in diabetic mice[7]. That is a laboratory measurement, not a demonstration that an oral supplement changes the course of proliferative retinopathy in a person.
How does this differ from the retinitis pigmentosa research?
The inherited degeneration work concerns photoreceptors dying from a genetic defect, and is covered on the retinal degeneration page. Diabetic retinopathy involves hyperglycaemia, inflammation and a leaking blood-retinal barrier, and the mouse work here reflects that: vascular permeability, leukocyte migration and TGR5 signalling[2]. Similar compound, different disease process.
References
- Tauroursodeoxycholic Acid Protects Retinal and Visual Function in a Mouse Model of Type 1 Diabetes. Pharmaceutics (2021).
- Tauroursodeoxycholic Acid Alleviates Endoplasmic Reticulum Stress-Mediated Visual Deficits in Diabetic tie2-TNF Transgenic Mice via TGR5 Signaling. Journal of Ocular Pharmacology and Therapeutics (2023).
- Protection of tauroursodeoxycholic acid on high glucose-induced human retinal microvascular endothelial cells dysfunction and streptozotocin-induced diabetic retinopathy rats. Journal of Ethnopharmacology (2016).
- Association of Serum Bile Acid and Unsaturated Fatty Acid Profiles with the Risk of Diabetic Retinopathy in Type 2 Diabetic Patients. Diabetes, Metabolic Syndrome and Obesity (2023).
- Association between Proliferative Diabetic Retinopathy and Serum Bile Acid Level in Patients with Type 2 Diabetes Mellitus. Endocrine, Metabolic and Immune Disorders Drug Targets (2021).
- Tauroursodeoxycholic acid protects retinal neural cells from cell death induced by prolonged exposure to elevated glucose. Neuroscience (2013).
- Ursodeoxycholic acid ameliorates diabetic retinopathy via reducing retinal inflammation and reversing the breakdown of blood-retinal barrier. European Journal of Pharmacology (2018).
These statements have not been evaluated by the Food and Drug Administration. This information is not intended to diagnose, treat, cure, or prevent any disease. Content is for informational purposes only and is not medical advice; consult a qualified healthcare provider before starting any supplement. As an Amazon Associate we earn from qualifying purchases.
