Dry Eye Research in 2021: 33 Notable Human Studies

The 2021 picture

What the research shows

The 2021 studies expanded dry-eye treatment options and measurement tools while showing again that symptoms and objective signs can point in different directions. Large datasets described disease burden and associations, although most could not establish which factor came first. Controlled comparisons and independent testing helped distinguish stronger findings from early leads.

Treatment trials

Controlled comparisons tested drops, gland procedures, and a computer reminder. Some treatments improved surface signs without meeting a main symptom endpoint; smaller or unmasked trials need replication before broad conclusions.

1. Lipid-based versus aqueous artificial tears

   A six-month, multicenter masked trial assigned 99 people with dry eye to lipid-based or aqueous artificial tears used at least four times daily. Symptoms improved in both groups by the first month, while tear breakup and staining improved later; lipid-layer grade rose only with lipid drops. The trial compared two active products in a modest sample, so it cannot estimate either product's benefit over no added treatment.

2. Water-free cyclosporine in the ESSENCE trial

   In a masked, vehicle-controlled trial, 328 patients used water-free cyclosporine 0.1% or vehicle twice daily for 12 weeks. Corneal staining improved more with cyclosporine by week four, but the trial did not meet its second main endpoint on the OSDI symptom score; a separate dryness rating favored treatment. The split result cautions against assuming an improved surface measure means equally clear overall symptom relief.

3. Donor serum versus a patient's own serum drops

   A double-blind crossover pilot gave patients with severe dry eye one month each of drops made from their own serum and from donor serum, with a washout between periods. Fifteen of 19 enrollees completed the study, and symptom changes were small and similar with both preparations. The small sample and short periods leave substantial uncertainty about equivalence and longer-term safety.

4. Adding gland expression to intense pulsed light

   Sixty patients with meibomian gland dysfunction were randomized to the order of intense pulsed light alone and light plus gland expression across four sessions. Adding expression improved tear breakup more than light alone, but symptom scores, staining, and gland secretion measures did not show a clear extra benefit. Both sequences received active light treatment, and the crossover design limits estimates of a durable stand-alone effect.

5. Reproxalap against vehicle in phase 2b

   A multicenter, double-masked trial assigned 300 dry-eye patients to two reproxalap concentrations or vehicle for 12 weeks. The higher concentration showed modest advantages on ocular dryness and nasal corneal staining, with a dose-response pattern. Several signs and symptoms were tested, so these phase 2 findings needed a later confirmatory trial.

6. Trehalose and hyaluronic acid after cataract surgery

   A randomized Italian study assigned 98 cataract patients with healthy preoperative ocular surfaces to trehalose plus hyaluronic acid, hyaluronic acid alone, or no additional drops. Tear production and symptom scores improved earlier in the combination group, and confocal nerve measures returned to baseline sooner over eight months. The single-center trial studied postoperative surface changes, so its findings do not directly establish benefit for chronic dry eye outside surgery.

7. Blink reminders for computer users

   A single-blind randomized trial compared frequent on-screen blink prompts with infrequent prompts in 46 computer users with dry eye. After 15 days, the frequent-prompt group had better OSDI symptom scores, and blink rate also improved in the study's follow-up comparison. The trial was small and short, and participants could see the prompting frequency, making expectation effects possible.

8. Topical estradiol in postmenopausal dry eye

   A masked phase 2 study randomized 104 postmenopausal women with moderate or severe dry eye to three estradiol dosing schedules or vehicle for three months. Tear production and symptoms improved within groups, but the main tear-production comparison did not separate estradiol from vehicle; only an inferior-staining comparison favored the highest dose. This exploratory result does not establish estradiol drops as effective treatment.

9. Intense pulsed light versus warm compress and massage

   A randomized trial assigned 132 patients with meibomian-related dry eye to intense pulsed light or warm compresses plus gland massage. The light group had more favorable short-term gland and tear measures and reported greater satisfaction at one month. The active treatments could not be masked easily, and the study's brief follow-up does not show whether the difference lasts.

10. Low-dose cyclosporine with trehalose

    A multicenter, masked phase 2 trial randomized 114 patients with moderate or severe dry eye among low-dose cyclosporine plus trehalose formulations, placebo, and a reference treatment. The main corneal-staining endpoint did not differ significantly in the full analysis; one formulation improved tear breakup on a secondary measure. The mixed results and small treatment arms do not support a firm efficacy claim.

11. Pulsed light plus gland expression versus warm compresses

    A randomized trial assigned 120 people with meibomian gland dysfunction to three sessions of intense pulsed light plus gland expression or warm compresses plus expression. Tear breakup and SPEED symptom scores improved more in the light-treatment group, although the staining comparison did not clearly differ. Without a sham light procedure and with brief follow-up, expectancy effects and durability remain open questions.

12. Short-course safety across four loteprednol trials

    A pooled secondary analysis combined four randomized trials of loteprednol etabonate 0.25% for dry-eye flares, comparing roughly 1,430 active-treatment participants with 1,438 vehicle recipients. Short-course adverse events and eye-pressure rises were uncommon, though numerically somewhat higher with the steroid. These are reused trial cohorts, not new patients; they excluded people with glaucoma and cannot establish safety with repeated or long-term use.

Measurement, imaging, and ocular-surface biology

These studies measured nerves, tear chemistry, immune cells, microbes, and gland images. Most identify promising patterns within selected samples; a useful clinical test also needs repeatability and validation in new patients.

13. Corneal nerve loss across dry-eye subtypes

    Confocal microscopy compared corneal nerves in 70 dry-eye patients and 42 controls. Nerve density and counts were lower in dry eye, especially in the aqueous-deficient subgroup compared with evaporative cases. This retrospective, cross-sectional comparison cannot show whether nerve changes preceded dryness or resulted from chronic surface disease.

14. Immune cells collected from the ocular surface

    Researchers used flow cytometry on ocular-surface washings from 48 dry-eye patients and 15 controls. Several leukocyte and T-cell populations differed, while greater neutrophil and lower natural-killer-cell proportions appeared in some aqueous-deficient eyes with staining. The modest cross-sectional groups show an immune pattern, not a validated marker or proof that changing those cells would improve symptoms.

15. Corneal dendritic cells as a possible immune clue

    A Veterans Affairs chart review counted activated dendritic cells on confocal images from 128 people with dry-eye symptoms. A threshold of two cells per central image detected systemic immune disease with 60% sensitivity and 77% specificity; counts fell in a 12-person before-and-after anti-inflammatory subgroup. The cutoff and treatment observation came from small, selected samples and need prospective validation.

16. Conjunctival gene activity in Sjögren dry eye

    Impression cytology compared conjunctival gene expression in seven women with Sjögren-related keratoconjunctivitis sicca and 19 controls. Fifty-three of 594 measured genes differed, with several immune-pathway signals also tracking tear instability or surface staining. The very small patient group and many gene comparisons make these mechanistic clues, not established treatment targets.

17. Ocular bacteria in diabetes and dry eye

    Tear samples from 115 people across diabetes-with-dry-eye, diabetes-only, dry-eye-only, and control groups underwent 16S sequencing. The combined diabetes and dry-eye group had the greatest measured bacterial diversity and a different community pattern. The groups were small and sampled once, so the study cannot tell whether microbes contributed to dryness or changed with it.

18. A thick lipid-layer reading is not always reassuring

    An interferometry study grouped 201 eyes from 102 dry-eye patients by tear lipid-layer thickness. Eyes with the thickest readings had worse tear breakup and surface staining than the middle group, despite similar meibomian dropout scores. A single lipid measurement did not capture overall tear health in this selected clinical sample, and analyzing two eyes per person complicates interpretation.

19. Repeatability of noninvasive tear breakup

    Eighty adult volunteers had repeated Keratograph noninvasive tear-breakup measurements. Results varied substantially within the same examiner and did not cleanly separate the dry-eye and other participants in this sample. The findings warn against relying on one reading, although they do not show that the test lacks value when combined with symptoms and other signs.

20. Tear metabolites distinguishing Sjögren syndrome

    A prospective study measured 104 tear metabolites in 40 newly diagnosed Sjögren patients and 40 people with other causes of dryness. A nine-metabolite model separated the groups with a reported area under the curve of 0.83. The small development dataset needs validation in an independent clinical population before the signature can guide diagnosis.

21. A portable tear-osmolarity test

    A retrospective comparison used the I-Pen device in 33 symptomatic dry-eye patients and 32 controls. Higher osmolarity tracked worse tear and symptom measures, and a cutoff chosen from these data showed about 91% sensitivity and 91% specificity. Selecting and testing the threshold in the same small clinic sample may overstate accuracy in routine care.

22. Automated reading of meibomian gland images

    Researchers trained and evaluated a deep-learning model on 728 anonymized clinical meibography images to outline glands and measure their shape. Agreement with manual outlines reached an average Dice score of about 0.84 on the test set. The images came from one device, and image segmentation accuracy does not by itself establish accurate dry-eye diagnosis or better patient outcomes.

Population patterns, impact, and related conditions

These studies combine large questionnaires, clinic cohorts, and administrative records. Their definitions differ, and most measure associations at one time rather than showing that an exposure caused dry eye.

23. Dry eye and physical and mental quality of life

    In the Dutch Lifelines cohort, 78,165 adults completed dry-eye and general health questionnaires. Dry eye was associated with poorer physical and mental quality-of-life scores even after adjustment for numerous other conditions, and more frequent symptoms tracked worse scores. The cross-sectional, self-reported measures cannot determine whether dry eye itself caused the difference.

24. Systemic illness and dry-eye severity in DREAM

    A secondary analysis followed 535 U.S. DREAM participants with moderate or severe dry eye using standardized eye tests and symptom scores. Sjögren syndrome, rosacea, rheumatoid arthritis, peripheral artery disease, and a daily smoking history accompanied worse objective signs, but none of the studied conditions clearly tracked symptom scores. These are observational comparisons within a treatment trial, so they do not prove the conditions caused worse eye findings.

25. Different risk patterns for dry-eye subtypes

    Researchers examined 1,125 visitors at a public science exhibition with questionnaires, tear tests, and meibography; 38% met their dry-eye criteria. Shorter sleep and female sex were among factors linked with aqueous deficiency, while screen exposure and contact lens wear were linked with evaporative disease. Exhibition visitors were a convenience sample, and one-time measurements cannot show that these exposures initiated either subtype.

26. Lifestyle factors measured alongside clinical dry-eye tests

    A community study examined 322 people using symptom questions and ocular-surface testing, finding that 111 met dry-eye diagnostic criteria. Greater reported screen time was associated with dry eye, while higher caffeine intake showed an inverse association after adjustment. One-time, self-reported habits leave room for confounding and cannot establish that changing either habit would prevent disease.

27. Sleep habits and dry eye in a large Japanese survey

    Baseline questionnaires from 106,282 Japanese adults linked frequent difficulty falling asleep with dry-eye diagnosis or severe symptoms; the adjusted odds were about twice as high at the greatest reported difficulty. Shorter sleep and waking tired showed similar patterns. Sleep and eye status were assessed together by report, so the direction of the relationship remains uncertain.

28. Medication patterns in a Dutch population cohort

    Among 79,606 Lifelines participants, researchers compared dry-eye symptoms with commonly used drug classes and individual medicines while adjusting for many underlying diagnoses. Associations persisted for proton pump inhibitors, glaucoma drugs, and anticholinergic medicines. Medication use can still stand in for illness severity or other unmeasured factors, so this survey cannot show that each drug causes dry eye.

29. Corneal injury in Taiwan claims records

    A population-based claims analysis matched people with aqueous-deficient dry eye to a comparison group and tracked corneal erosions, ulcers, or scars across records from 1997–2013. Recorded corneal damage was more frequent in the dry-eye group, with a hazard ratio of 2.70. Claims codes do not measure the severity of every eye at baseline, and residual differences between groups could affect the estimate.

30. Air pollution and reported dry eye among Taiwanese women

    Investigators analyzed 27,605 Taiwan Biobank participants, assigning pollution estimates from residential municipalities and comparing them with reported dry eye. Higher nitrogen dioxide exposure and temperature were associated with dry eye among women after adjustment for selected factors. Area-level exposure and the cross-sectional design limit individual causal interpretation despite the long span of environmental data.

31. Dry-eye burden in a United Kingdom survey

    An online survey compared 1,002 adults who reported dry eye with 1,003 without it using vision and daily-activity questionnaires. The dry-eye group reported poorer vision-related functioning, and greater symptom severity tracked more work and activity problems. Recruitment by email and self-reported diagnosis make the findings vulnerable to selection and reporting bias.

32. Dry-eye prevalence and associated conditions in the Lifelines cohort

    Questionnaires from 79,866 adults in the Dutch Lifelines cohort classified 9.1% as having dry eye. Female sex, contact lens wear, and a wide range of systemic conditions were associated with the condition, providing a large population reference for later analyses of the same cohort. The estimate uses reported symptoms and prior diagnosis rather than a standardized eye examination, and the cross-sectional design does not establish causes.

33. Dry-eye symptoms in a Ghanaian community survey

    A stratified community survey questioned 1,316 adults in Ghana and found a high proportion above its symptom-score threshold; older age, female sex, arthritis, and ocular allergy were associated with those scores. The survey measured symptoms rather than confirming dry-eye disease with tear or surface tests. Its cross-sectional associations cannot establish causes, and the symptom figure should not be read as clinical disease prevalence.