Keratoconus is a progressive ectatic corneal disorder in which the cornea thins, steepens, and bulges into a cone-like shape, producing irregular astigmatism and progressive visual impairment. Once viewed primarily as a structural abnormality, the disease is now understood as a multifactorial condition involving genetic, environmental, biomechanical, and cellular mechanisms. Alterations in stromal collagen organization and cross-linking, inflammatory signaling, and oxidative stress have all been implicated in its development and progression, while behavioral factors such as eye rubbing and atopy have been independently associated with the disease.

That shifting understanding has pushed the ocular surface itself into the spotlight. Researchers have documented elevated tear concentrations of inflammatory mediators such as MMP-9 and HMGB1 alongside reduced levels of the antioxidant SOD3 in keratoconus, painting a picture of an altered inflammatory and oxidative microenvironment at the eye’s surface. In vivo confocal microscopy has revealed reduced corneal nerve density, disorganized and increasingly tortuous subbasal nerve plexuses, and altered corneal sensitivity. Quantitative meibography has additionally identified structural abnormalities of the meibomian glands, the lipid-secreting structures that stabilize the tear film, with gland morphology linked to lipid-layer thickness and disease severity.

To bring order to this scattered evidence, the research team systematically searched PubMed, Embase, Web of Science, and the Cochrane Library from inception through September 1, 2026. They included comparative studies reporting at least one of four key outcomes: the OSDI questionnaire, fluorescein tear break-up time (TBUT), noninvasive tear break-up time (NITBUT), or the Schirmer I test (SIT), a measure of tear secretion. Meibomian gland dysfunction and meibomian gland atrophy were captured as secondary findings when reported. Two reviewers independently screened the 642 records initially identified, ultimately including seven studies published between 2019 and 2026 across six countries or regions, with designs ranging from cross-sectional surveys to prospective and retrospective comparative studies.

A distinguishing feature of the meta-analysis was its unusually strict handling of a problem that plagues ophthalmology research: the eye. Because most people have two eyes, and both eyes of one person are statistically correlated, treating each eye as an independent data point can artificially inflate sample sizes and produce spurious significance. The researchers required that any eye-specific outcome either use one eye per participant or apply statistical methods that properly account for inter-eye correlation before pooling data. When these conditions could not be verified, they declined to calculate a pooled estimate, summarizing study-level findings instead.

The results revealed a striking asymmetry in evidence quality. Five studies contributed participant-level OSDI data, and the pooled analysis showed keratoconus patients scoring 19.45 points higher than controls, with a 95 percent confidence interval of 12.27 to 26.63 points. Substantial heterogeneity was present, with an I-squared of 79.9 percent, and the 95 percent prediction interval ranged from 3.62 to 35.29 points, suggesting considerable variability in how large the symptom gap might be in any future study. Leave-one-out analyses confirmed the finding was not driven by any single investigation, with pooled estimates remaining between 17.6 and 21.1 points regardless of which study was excluded.

The picture for objective tear-film measures was far murkier. Fluorescein TBUT, which uses dye to detect when the tear film ruptures, was shorter in keratoconus patients by 2.74 seconds on average, but the confidence interval spanned from 7.81 seconds shorter to 2.32 seconds longer, and the I-squared of 93.8 percent indicated extreme between-study inconsistency. All three studies reporting NITBUT, a gentler measurement that avoids dye-induced tear disruption, found lower values in keratoconus, but two of the three failed to clearly specify how eyes were selected or aggregated, precluding a pooled estimate. Schirmer test findings varied across the five studies reporting it, with some showing clear deficits in keratoconus and others showing little or no difference.

The categorical outcomes added another layer of complexity. In one study, the prevalence of diagnosed dry eye disease did not differ significantly between keratoconus and control eyes, yet meibomian gland dysfunction was significantly more common in keratoconus eyes, affecting 22 of 33 eyes compared with 8 of 27 controls. A French case-control study by Morice and colleagues found abnormal meibomian gland atrophy in 64.8 percent of keratoconus eyes versus 27.3 percent of control eyes, with keratoconus conferring nearly fivefold higher odds after adjustment. These gland-level findings suggest that structural changes in the oil-producing apparatus of the eye may be a more consistent feature of keratoconus than conventional dry eye metrics capture.

The authors caution that OSDI, while robust, is not specific to dry eye disease and may also reflect visual fluctuation, irregular corneal optics, contact lens intolerance, and general ocular irritation, all of which are common in keratoconus independent of tear film dysfunction. Subgroup analysis examining whether control groups explicitly excluded people with dry eye or ocular surface disease found no statistically significant difference in the OSDI effect, though the small number of studies limits the power of that comparison. The methodological quality of the included studies, assessed with an adapted Newcastle-Ottawa Scale, ranged from 4 to 7 out of a possible 9 stars, reflecting variability in control selection and comparability adjustments.

Several plausible biological mechanisms could explain the ocular surface findings. Repetitive eye rubbing may promote epithelial injury and extracellular matrix remodeling, while elevated tear levels of proinflammatory cytokines including IL-1β, IL-6, TNF-α, and matrix metalloproteinases may drive epithelial dysfunction and stromal degradation. Tear proteomic studies further point to disrupted inflammatory, extracellular matrix, and epithelial homeostatic pathways. Conjunctival cytologic abnormalities and impaired tear function have also been observed. Yet the evidence remains predominantly observational, leaving open whether these ocular surface changes are causes, consequences, or parallel manifestations of the corneal ectasia itself. The researchers call for larger prospective multicenter studies with standardized dry eye definitions, clearly specified analysis units, and harmonized tear-testing protocols to resolve the inconsistencies that currently cloud the objective tear-film picture in keratoconus.

Subject of Research: Ocular surface symptoms and tear-film parameters in keratoconus assessed through systematic review and meta-analysis

Article Title: Ocular surface symptoms and tear-film parameters in keratoconus: a systematic review and meta-analysis

Article References: Cao, Y., Yang, Y., Deng, W., & Yang, N. (2026). Ocular surface symptoms and tear-film parameters in keratoconus: a systematic review and meta-analysis. Frontiers in Medicine, 13, Article 1797492. https://doi.org/10.3389/fmed.2026.1797492

Image Credits: AI Generated

DOI: 10.3389/fmed.2026.1797492

Keywords: keratoconus, dry eye disease, ocular surface, tear film, OSDI, meta-analysis, tear break-up time, Schirmer test, meibomian glands, corneal ectasia, ophthalmology, systematic review