Dry Eye Research in 2025: 31 Notable Human Studies

The 2025 picture

What the research shows

In 2025, controlled studies tested several practical and experimental treatments, while early trials raised questions for larger follow-up studies. Population research showed that dry-eye burden varies by setting. Diagnostic work reinforced the value of combining symptoms with tear and surface findings, since individual markers and image models could not yet capture the full condition.

Controlled treatments and early clinical experiments

The designs here range from multisite randomization to very small first-in-human work. A novel biological signal is not the same as an established treatment.

1. Exosome drops in Sjögren dry eye: a very small trial

   An eight-person randomized trial put mesenchymal-stem-cell-derived exosome drops in one eye and saline in the other. Several eye-level tear and staining measures favored the treated eyes. The reported symptom comparison used a whole-person questionnaire, so it cannot show an eye-specific benefit. This tiny, short trial cannot establish effectiveness or long-term safety.

2. Omega-3 after cataract surgery in gland dysfunction

   An investigator-masked multicenter trial randomized 107 people with meibomian gland dysfunction after cataract surgery to re-esterified triglyceride omega-3 or control. At 12 weeks, tear breakup, staining, and symptom questionnaires favored supplementation, while overall gland measures did not clearly differ. This selected postoperative population and short course differ from broader omega-3 trials.

3. Low-level light alone versus light plus IPL

   A double-masked paired-eye trial assigned the eyes of 24 patients with gland-related dry eye to low-level light alone or combined low-level and intense-pulsed-light treatment. Both regimens improved some symptoms and gland measures from baseline, with mixed differences at later follow-up. Whole-person symptoms are difficult to attribute to one treated eye, and the sample was small.

4. Fifty-percent autologous serum for severe Sjögren dry eye

   A double-masked contralateral-eye trial compared 50% autologous serum with artificial tears in ten patients with severe Sjögren dry eye over three months. Tear stability, tear production, visual acuity, and corneal-nerve measures improved on the serum-treated side; nerve changes correlated with less staining. Bacteria detected in used serum bottles underscored handling concerns, and the ten-person trial cannot define broad safety or lasting benefit.

5. Topical androgen gel: a pilot with a safety signal

   A randomized pilot assigned 24 patients to eyelid testosterone gel or vehicle for four weeks. Some symptom, tear, and meibum measures improved in the active group, but several other dry-eye measures did not. Serum testosterone rose markedly in women receiving active gel, making dose and systemic exposure important questions before a larger trial.

6. Platelet-rich plasma against conventional tears

   A prospective randomized study assigned 40 people with moderate to severe dry eye to 20% autologous platelet-rich plasma or conventional artificial tears for three months. Symptoms, tear measures, staining, and impression cytology favored plasma. The study was small, and the abstract does not describe masking, so expectancy and other care may influence reported benefit.

7. Extra lubrication after cataract surgery missed significance

   An investigator-masked randomized study stratified cataract patients by ocular-surface risk and added preservative-free lubricant for high-risk patients. Among 83 analyzed participants, some three-month symptom and tear measures trended favorably, but no statistically significant between-group difference was found. The result does not establish that the added regimen prevents postoperative dry eye.

8. Blink exercises added to artificial tears

   A randomized study assigned 100 dry-eye participants to artificial tears alone or tears plus three days of repeated blinking exercises. Exercise recipients had better reported symptoms, blink completeness, and tear-breakup measures than controls. The brief intervention and follow-up leave persistence of the benefit unknown.

9. Aloe vera and hypromellose drops: paired-eye test

   A double-masked contralateral-eye study randomized one eye of each of 20 dry-eye patients to Aloe vera–hypromellose lubricant and the other to saline for three months. Tear MMP-9 and IL-6, tear breakup, and some symptoms improved on the active side. The tiny eye-level comparison and shared whole-person symptoms limit the clinical conclusion.

10. Two cyclosporine strengths in Sjögren dry eye

    A retrospective comparison followed 30 Sjögren patients using 0.1% cyclosporine and 30 using 0.05%. Several signs and symptoms improved more with the higher strength, but treatment discontinuation was much more frequent. Nonrandomized treatment selection and the short three-month follow-up mean the study cannot settle the best concentration for an individual patient.

Prevalence, associated conditions, and daily burden

These studies use surveys, registries, exposure estimates, or clinical samples. They describe their populations and cannot by themselves identify causes.

11. A population survey across Spain

    A telephone survey of 3,019 Spanish adults estimated dry eye at 16.6% using Women's Health Study criteria and 22.5% using another symptom definition; 12.3% reported a clinical diagnosis. Women reported more disease, and many younger adults reported symptoms without a diagnosis. The contrasting estimates show why definition matters.

12. Quality of life in a dry-eye registry

    A registry analysis of 368 patients examined quality-of-life questionnaires using Rasch methods. Mixed aqueous and evaporative disease and neuropathic corneal pain were associated with worse reported function than comparison subgroups, while relationships between signs and symptoms were weak. The clinic registry is descriptive and cannot establish that a subtype caused the difference.

13. Pollution, temperature, and clinic visits

    A time-stratified case-crossover analysis in Hangzhou, China, linked daily dry-eye outpatient visits with air pollutants and temperature from 2014 to 2019. Several pollutant levels tracked more visits, with some stronger associations in cooler conditions. Citywide exposure and visit counts are indirect measures; they do not prove what caused a patient's symptoms.

14. Neighborhood green space and dry-eye measures

    A Chinese case-control analysis compared 450 dry-eye patients with 900 controls, and a separate community sample measured tear meniscus in 140 people with diabetes. Greater long-term neighborhood green space was associated with lower dry-eye odds and a larger tear meniscus. Geographic and social differences may contribute despite statistical adjustment.

15. Ocular problems during pediatric dupilumab therapy

    A prospective registry followed 104 children receiving dupilumab for atopic dermatitis for a median of about 70 weeks. Roughly one-third developed ocular-surface symptoms, and some needed anti-inflammatory eye treatment; high baseline IgE was associated with risk. This treated cohort informs monitoring but does not compare dupilumab with an untreated control.

16. Type 1 diabetes, neuropathy, and the ocular surface

    A cross-sectional comparison assessed 97 people with type 1 diabetes and matched controls, with nerve testing within the diabetes group. Dry-eye and gland measures were worse with diabetes, especially among those with peripheral neuropathy. Age, disease duration, and glucose control also tracked outcomes, so the study cannot isolate neuropathy as the sole cause.

17. Smoking and corneal nerve measurements

    A cross-sectional study compared 250 smokers with 272 nonsmokers using confocal nerve imaging and ocular-surface tests. Smokers had lower nerve density and more inflammatory-cell findings, with some measures related to smoking duration. The study supports an association but cannot rule out other differences between the groups.

18. Preserved versus preservative-free artificial tears in a large survey

    A survey of 67,951 respondents included 8,039 artificial-tear users. People using benzalkonium-chloride-preserved drops reported somewhat more itching and blurred vision than preservative-free users after adjustment. Treatment choice may reflect underlying severity, and the cross-sectional data cannot establish that preservative exposure caused symptoms.

19. Meibomian gland structure in children with higher weight

    Investigators imaged the eyelid glands of 200 children aged 10 to 18. Higher body-mass-index ranking was associated with more gland atrophy or tortuosity, while shorter tear breakup tracked gland loss and symptoms. The observational study cannot determine whether weight, inflammation, or another factor drives the gland changes.

20. Dry-eye diagnoses across South Korean insurance records

    A nationwide South Korean insurance sample covered more than 6.3 million people from 2016 through 2020 and identified roughly 567,000 with a recorded dry-eye diagnosis. The five-year coded prevalence was 8.9%, higher among women and older adults, and hyaluronic-acid drops were frequently prescribed. Diagnosis and prescription claims measure treated care, not everyone with symptoms or the effectiveness of the drops.

Blinking, nerves, tear markers, and diagnostic tools

Several large method studies show that a result depends on the disease definition and how the test is performed. Promising markers still need independent validation in real diagnostic settings.

21. Incomplete blinks versus blink rate

    An investigator-masked cross-sectional study assessed 453 community residents. More incomplete blinking was associated with dry-eye classification, unstable tears, and gland findings; blink rate itself showed no clear association. The proposed 40% incomplete-blink threshold was derived in this sample and needs testing elsewhere.

22. InflammaDry performance changes with the reference definition

    A retrospective analysis of 1,515 patients compared a tear MMP-9 point-of-care result with two clinical dry-eye definitions. Reported sensitivity dropped from about 81% under stricter trial-style criteria to 70% under broader consensus criteria, while specificity in the selected comparison was low. The test identifies inflammation, not every dry-eye case.

23. What is missed when a dry-eye sign is omitted?

    Across five sites, researchers applied a full diagnostic battery to 1,427 people, then recalculated classifications as individual tear or staining tests were removed. Relying on one homeostatic sign or corneal staining alone missed many people who met the full rule. The analysis supports a combined assessment, though its reference standard is the algorithm being tested.

24. Corneal sensory changes after LASIK

    A prospective cross-sectional comparison grouped 89 post-LASIK patients into neuropathic pain with dry eye, dry eye alone, or asymptomatic controls. The pain group reported more severe symptoms and more nonocular sensitization features, alongside corneal nerve and sensitivity differences. These selected groups cannot establish a single diagnostic cutoff.

25. Blood inflammatory proteins in a large twin cohort

    A hypothesis-free analysis measured 74 serum inflammatory proteins in 2,602 TwinsUK participants. A small set, including CXCL9 and CXCL10, remained associated with questionnaire-defined dry eye after multiple-testing correction. The predominantly female cohort and cross-sectional associations do not make any protein an established blood test for dry eye.

26. Young adults followed for one year

    A prospective study followed 50 adults aged 18 to 25, measuring surface signs, tear proteins, and cytokines. Redness, lid-wiper changes, and blink rate worsened modestly over a year; several tear markers differed with evaporative dry eye. The small sample and many measured markers make this an exploratory view of early disease.

27. Noninvasive versus fluorescein tear-breakup time

    A cross-sectional examination of 132 Norwegian adults aged 65 compared two noninvasive tear-breakup outputs with fluorescein breakup. The first noninvasive breakup occurred sooner on average than the fluorescein result, while the device's average breakup time was similar. The methods are not interchangeable simply because they share the word 'breakup.'

28. Imaging and tear proteins in neuropathic corneal pain

    A comparison of 54 neuropathic-pain patients, 53 dry-eye patients, and 54 controls combined confocal imaging, pain scores, and tear proteomics. Corneal microneuroma features and selected proteins differed across groups, and an internally derived combination showed high discrimination. Independent validation is needed before using it to classify an individual with eye pain.

29. Quantitative MMP-9 from a Schirmer strip

    A method-development study tested a fluorescent MMP-9 assay on Schirmer strips in 121 dry-eye patients and 41 controls. At a study-derived cutoff, sensitivity was about 74% and specificity about 78%, and levels changed with treatment response. Threshold selection in the same sample may overstate performance; external clinical validation is required.

30. AI analysis of tear-breakup videos

    An image-analysis study trained a model on fluorescein tear-breakup videos from 143 eyes, including 98 with dry eye. It reported high accuracy for breakup detection and pattern classification in its dataset. The small development sample and lack of broad external validation limit claims about routine diagnostic use.

31. Tear MMP-3 as an exploratory marker

    A cross-sectional human comparison measured tear MMP-3 in 59 dry-eye patients and 21 controls, alongside laboratory experiments. Patient levels were higher and correlated with corneal staining and reported disease duration; a cutoff performed well within the study. The marker and threshold need independent testing before clinical use.