The Gut-Eye Axis and Dry Eye: Plausible, Not Proven

Dry eye may involve more than a thin tear film or clogged oil glands. A growing body of research proposes that changes in the gut microbiome can alter immune signaling throughout the body and eventually inflame the lacrimal glands and ocular surface. It is a provocative idea: a gut-targeted treatment could address biology that artificial tears cannot reach.

The clinical evidence has not caught up. The only randomized controlled trial described here enrolled 41 people with dry eye disease and found no statistically significant treatment-period change in its three inflammatory biomarker endpoints after four months of probiotic and prebiotic supplementation. The gut-eye axis is a credible hypothesis, not an established dry-eye treatment.

How the gut could affect the eye

The proposed pathway centers on immune balance. In a healthy intestine, microbial communities help regulate inflammatory T-helper 17 cells (Th17 cells) and regulatory T cells (Tregs). Th17 cells produce interleukin-17 (IL-17), a cytokine involved in inflammatory responses; Tregs help restrain excessive inflammation.

A 2025 review by Wang et al. describes how microbial antigens, including those associated with Klebsiella pneumoniae and segmented filamentous bacteria, can activate intestinal Th17 cells. In the proposed model, dysbiosis weakens Treg restraint, allowing IL-17-related inflammation to extend beyond the gut. The review largely discusses immune-mediated eye disease and does not establish that this sequence causes dry eye in people.

A 2024 dry-eye-focused review by Song et al. applies that framework to the ocular surface. It proposes that dysbiosis may reduce short-chain fatty acid production and disrupt mucosal immunoglobulin A secretion, while increasing systemic exposure to lipopolysaccharide and pro-inflammatory cytokines. These signals could impair the lacrimal functional unit: the linked tissues, nerves, glands, and tear-film components that maintain a stable ocular surface.

This is biologically coherent, but coherence is not proof. Dry eye has several drivers, including tear evaporation, meibomian gland dysfunction, medications, autoimmune disease, screen exposure, and aging. A microbiome association can sit within that picture without becoming the root cause of every scratchy eye.

Why MMP-9 matters—and why it may be a blunt test

Matrix metalloproteinase-9 (MMP-9) is a protease often elevated in dry eye disease. It can damage the glycocalyx of the tear film and epithelial tight junctions on the ocular surface. In the gut-eye model, however, MMP-9 sits far downstream: gut microbial changes would first need to alter immune regulation, then systemic inflammatory signaling, then lacrimal or ocular-surface biology before changing tear MMP-9.

That makes MMP-9 relevant to dry-eye inflammation, but imperfect as a test of a gut intervention. Serum C-reactive protein (CRP) poses a related problem. CRP responds well to larger-scale systemic inflammation and may miss modest mucosal immune shifts. A study that measures neither microbiome changes nor closer pathway markers such as IL-17, fecal immunoglobulin A, or short-chain fatty acids cannot show whether a supplement changed the proposed upstream biology.

A 2026 systematic review by Priyanka et al. found that dysbiosis has been associated with several ocular disorders, including ocular-surface disease. Its breadth supports the idea that gut-derived immune effects may reach the eye. It does not demonstrate that changing the gut microbiome improves dry-eye symptoms or signs.

The one controlled trial found no biomarker benefit during treatment

The most direct evidence comes from a 2026 double-masked randomized controlled trial by Tavakoli et al. The trial assigned 41 patients with dry eye disease to oral probiotic plus prebiotic supplementation (23 participants) or control (18 participants) for four months. The reported endpoints were tear MMP-9, tear tissue inhibitor of metalloproteinases-1 (TIMP-1), and serum CRP.

During supplementation, neither group showed a statistically significant change in MMP-9, TIMP-1, or CRP; all reported p values were greater than 0.05. The available abstract does not provide confidence intervals or effect estimates, so it cannot tell readers how large or small the between-group differences were. It supports a narrower conclusion: this formulation, in this small trial, did not produce statistically significant changes in those biomarkers over four months.

The abstract also does not report the patient-centered outcomes that would matter most in practice: Ocular Surface Disease Index (OSDI) symptom scores, tear-breakup time, Schirmer testing, or corneal staining. A biomarker result cannot substitute for a symptom result, and a null biomarker result cannot prove that symptoms did not change. The question remains unanswered by the published abstract.

The post-treatment rise is interesting, not evidence of a rebound effect

One month after supplementation stopped, MMP-9 rose significantly in the treatment group (p < 0.05). The MMP-9:TIMP-1 ratio also increased significantly in both groups (p < 0.05). This result invites a tempting story: perhaps supplementation partly suppressed ocular-surface inflammation, then stopping it allowed inflammation to return.

That is possible, but the data do not establish it. With 23 treated participants and 18 controls, chance findings and regression to the mean are real concerns, particularly when several biomarkers and time points are assessed. The intervention might have had a biological effect that the treatment-period measurements missed. The finding might also be random variation, or it could reflect a change unrelated to beneficial suppression during treatment. Without microbiome sequencing, immune-marker kinetics, and replication, calling this a therapeutic rebound would outrun the evidence.

Why probiotics elsewhere do not validate probiotics for dry eye

Probiotics and related interventions can affect some non-eye outcomes, but effects depend on the condition, product, population, and endpoint. A 2025 network meta-analysis in inflammatory bowel disease reported that probiotics and fecal microbiota transplantation outperformed placebo across the outcomes evaluated, while a 2022 meta-analysis in type 2 diabetes reported statistically significant changes in glucose-homeostasis markers with microbiome-modulating interventions. Those results show that gut-targeted treatments can have measurable effects in some contexts. They do not identify a dry-eye strain, dose, duration, or clinical benefit.

That distinction matters because probiotic effects are strain-specific. The available abstract of the Tavakoli trial does not identify the organisms, prebiotic compounds, or doses used. A label that says "probiotic" is not a pharmacologic mechanism of action. It is a category with a very crowded family reunion.

What a useful next trial would measure

A stronger test would enroll enough participants to detect a clinically meaningful difference and prespecify clinical co-primary endpoints such as OSDI and tear-breakup time. It should identify the strains, doses, and prebiotic ingredients; measure adherence; and compare microbiome and immune markers before, during, and after treatment. IL-17, fecal immunoglobulin A, short-chain fatty acids, and stool sequencing would test whether the proposed pathway moved at all. Planned post-cessation follow-up could determine whether the MMP-9 signal repeats.

For now, standard dry-eye care has a stronger evidentiary footing than probiotics targeted at the gut-eye axis. People with persistent burning, fluctuating vision, light sensitivity, or eye pain should seek an eye-care evaluation rather than swap proven treatment for a supplement. Oral probiotics may be reasonable for an unrelated indication when a clinician agrees, but current evidence does not support presenting them as a dry-eye therapy.

The gut-eye axis has credible immunology behind it. The clinical claim still rests on one small randomized trial with null treatment-period biomarker results and no reported symptom outcomes in the available abstract.