Abstract
PURPOSE:
Dry eye disease (DED) is a common ocular surface disorder associated with ocular discomfort, visual disturbance, and reduced quality of life. Vitamin D has anti-inflammatory and immunomodulatory properties, and deficiency may predispose individuals to ocular surface disease. This study evaluated the association between vitamin D deficiency and incident DED.
DESIGN:
Retrospective cohort study.
SUBJECTS:
Adults aged ≥18 years identified from the TriNetX U.S. Collaborative Network. After 1:1 propensity score matching, 6,047,502 patients with vitamin D deficiency were compared with 6,047,502 matched controls without vitamin D deficiency.
METHODS:
De-identified electronic health records with up to 20 years of follow-up were analyzed. Vitamin D deficiency was defined using ICD-10 code E55. Controls were patients undergoing routine medical examinations without abnormal findings (ICD-10: Z00.00). Propensity score matching was performed for age, sex, and relevant comorbidities. Time-to-event analyses were conducted using Cox proportional hazards models. Kaplan–Meier curves and log-rank tests compared cumulative incidence between cohorts. Risk ratios (RRs) and odds ratios (ORs) were calculated from end-of-follow-up incidence proportions.
MAIN OUTCOME MEASURES:
Incident dry eye disease, defined by ICD-10 code H04.12.
RESULTS:
DED developed in 196,639 patients (3.3%) with vitamin D deficiency and 160,141 patients (2.7%) without deficiency. Vitamin D deficiency was associated with a significantly increased risk of DED (hazard ratio [HR], 1.286; 95% CI, 1.277–1.294; p < .001). Kaplan–Meier curves demonstrated a higher cumulative incidence in the vitamin D–deficient cohort (log-rank χ2 = 5590.618; p < .001). The risk ratio was 1.242 (95% CI, 1.234-1.250; p < .001), and the odds ratio was 1.250 (95% CI, 1.242-1.259; p < .001). Limitations include reliance on diagnostic coding and lack of serum vitamin D level data.
CONCLUSIONS:
Vitamin D deficiency was significantly associated with an increased risk of developing DED, with a 28.6% higher hazard compared with matched controls. These findings suggest that vitamin D deficiency may represent a modifiable risk factor for DED and support further prospective studies to evaluate the role of vitamin D supplementation in prevention and management.
INTRODUCTION
Dry eye disease (ded) is a heterogenous ocular surface condition characterized by a loss of homeostasis and associated ocular symptoms. Tear film instability, hyperosmolarity, ocular surface inflammation and damage and neurosensory abnormalities play etiological roles.1 A meta-analysis of United States (US) studies estimated a DED prevalence of approximately 8.1% (95% confidence interval (CI) 4.9-13.1%) in the adult population, acknowledging that different diagnostic definitions (symptoms and/or signs) were used across studies.2 Despite differences in prevalence reports, DED is widely acknowledged as a common condition that drives a large proportion of eye clinic visits and imposes a substantial burden on patients and the healthcare system.2–4 Functionally, ocular pain symptoms coupled with fluctuating vision can impact the ability to perform activities of daily living and significantly reduce quality of life.5,6 Current DED therapies target multiple facets of disease, including tear supplementation with ocular lubricants, anti-inflammatory treatment with limited-duration topical corticosteroids or immunomodulators such as cyclosporine, and Meibomian gland dysfunction (MGD)-directed therapies including lid/physical measures and oral macrolide or tetracycline antibiotics; dietary omega-3/polyunsaturated fatty acids have also been investigated as adjunctive options.7 However, these therapies vary in effectiveness and are often associated with burdensome side effects.8 The chronic burden of DED and the limited efficacy of existing treatments highlight the need for effective preventive and new therapeutic strategies.
Vitamin D is a multifunctional hormone whose main physiological role is to maintain calcium and phosphate balance, mineral metabolism, and skeletal health.9 It also has important anti-inflammatory and immune-modulating effects, mediated through regulation of gene transcription, and it can reduce inflammatory signaling and influence programmed cell death.9 Given that DED is multifactorial with inflammation as an important component, vitamin D may serve as a potential mediator in its pathogenesis.10
Several studies have explored the relationship between vitamin D and DED with mixed findings. Some have found that vitamin D deficiency impacts symptoms and signs of DED. In a prospective clinical study of 34 individuals with serum vitamin D deficiency and 21 controls with normal vitamin D levels, the case group demonstrated significantly lower tear film break-up time (5.18 ± 2.15 vs 7.36 ± 3.10 seconds; p = .01) and Schirmer test scores (12.18 ± 6.44 vs 18.57 ± 8.99 mm; p = .007) compared to controls, supporting an association between vitamin D deficiency and impaired tear film stability and production.11 Vitamin D deficiency has also been implicated in DED symptoms. A 2020 systematic review and meta-analysis included 10 clinical studies (18,919 participants).12 Pooling 5 studies with DED as the exposure, heterogeneously defined in each study by a mix of symptoms and/or signs, and 5 studies with serum vitamin D as the exposure, the authors found that patients with DED had mean serum 25-hydroxyvitamin D levels 3.99 ng/mL lower than controls.12 In subgroup meta-analysis restricted to 5 studies that used vitamin D deficiency as the exposure, individuals with vitamin D deficiency had OSDI scores 10.70 points higher and Schirmer test values 6.38 mm/5 min lower than controls, while tear break-up time was similar between the groups, indicating that vitamin D deficiency is associated with a more severe DED phenotype across symptoms and signs.12 On the other hand, other studies did not find relationships between vitamin D and DED. A cross-sectional study of 740 Korean adults found no significant relationship between serum vitamin D levels and OSDI-measured dry eye symptoms.13 A similar pattern was observed in a study of 247 US male Veterans, where serum vitamin D levels showed no relationship with tear parameters.14 Taken together, these studies suggest that vitamin D deficiency may influence certain aspects of DED, although results across the literature remain inconsistent.
Vitamin D supplementation has been investigated as a treatment for dry eye disease, with predominantly beneficial effects, particularly among individuals with vitamin D deficiency. In a 2024 systematic review and meta-analysis of eight studies including 439 participants, vitamin D supplementation was associated with improved symptoms on the OSDI, with a standardized mean difference (SMD) of −1.10 (95% confidence interval −1.45 to −0.74), and improved clinical signs, including increased tear production on the Schirmer tear secretion test (SMD 1.43, 95% confidence interval 0.81 to 2.05) and improved tear film stability measured by tear film break-up time (SMD 1.19, 95% confidence interval 0.83 to 1.55), supporting a potential therapeutic role for supplementation.15 Similar findings were noted in smaller studies. In a retrospective observational study, 105 patients with DED refractory to conventional treatment and vitamin D deficiency or insufficiency (mean serum 25-hydroxyvitamin D level 10.52 ± 4.61 ng/mL) received a single intramuscular injection of 200,000 international units of cholecalciferol. Several tear film parameters improved, at least transiently, including tear stability (3.16 ± 2.27 → 5.19 ± 2.34 seconds at 6 weeks, p = .001) and production (6.69 ± 3.92 → 8.92 ± 7.60 millimeters at 6 weeks (p = .015).16 Taken together, these findings suggest that vitamin D supplementation may serve as a useful adjunct therapy for DED, but larger, high-quality randomized trials are still needed to define optimal dosing, durability of effect, and the patient subgroups most likely to benefit.
Overall, although significant associations between vitamin D deficiency and DED have been reported, the literature is limited by small sample sizes, cross-sectional designs, and methodological inconsistencies across studies. Large-scale, longitudinal analyses are thus needed to clarify the strength and direction of this relationship. This study aims to address these gaps through a retrospective cohort study using a large, federated database to evaluate the association between vitamin D deficiency and incident DED.
METHODS
This retrospective cohort study utilized data from the TriNetX US Collaborative Network, a federated research platform that aggregates de-identified electronic health records from more than 180 healthcare organizations across the United States. The platform is fully compliant with the Health Insurance Portability and Accountability Act (HIPAA), with all patient-level data de-identified in accordance with Section 164.514(a) of the HIPAA Privacy Rule. TriNetX’s integrated analytic tools, including incidence estimation, outcomes analyses, and propensity score matching, allow patient-level analyses while reporting only aggregate-level results, thereby preserving data confidentiality. Because this study relied exclusively on de-identified data, it did not constitute human subjects research, and review by an Institutional Review Board was not required. The study adhered to the ethical principles outlined in the Declaration of Helsinki and its subsequent amendments and was conducted in accordance with the STROBE (STrengthening the Reporting of Observational Studies in Epidemiology) guidelines.
The TriNetX database was queried for records spanning September 2005 through September 2025. Eligible participants were adults aged 18 years or older at the time of their most recent qualifying clinical encounter. Two cohorts were defined using International Classification of Diseases, Tenth Revision (ICD-10) codes. Within the TriNetX platform, historical ICD-9 diagnoses are algorithmically mapped to their ICD-10-CM equivalents, allowing cohort definition using ICD-10 codes over the entire study period. Cohort 1 consisted of patients with a documented diagnosis of Vitamin D deficiency (ICD-10: E55), while Cohort 2 comprised patients who had undergone routine adult medical examinations without abnormal findings (ICD-10: Z00.00) and who had no prior history of vitamin D deficiency. The index date was defined as the first recorded encounter for vitamin D deficiency in Cohort 1 and the first documented routine adult examination in Cohort 2. Encounters occurring before September 2005 were excluded. Outcomes were assessed beginning one day after the index event, and all subsequent events recorded through September 2025 were included in the observation period.
To reduce baseline imbalances and minimize confounding, propensity score matching was performed at a 1:1 ratio using the nearest-neighbor method without replacement. Matching was restricted to 5 prespecified variables: age at index and sex in the demographics category; ischemic heart disease (ICD-10: I20–I25) and diabetes mellitus (ICD-10: E08–E13) in the diagnosis category; and body mass index (BMI) in the laboratory category. Other characteristics were captured and reported in the analysis, including race and ethnicity, as well as clinical diagnoses relevant to ocular health such as menopausal and perimenopausal disorders (ICD-10: N95), essential (primary) hypertension (ICD-10: I10), nicotine dependence (ICD-10: F17), other hypothyroidism (ICD-10: E03), thyrotoxicosis/hyperthyroidism (ICD-10: E05), depressive episode (ICD-10: F32), and generalized anxiety disorder (ICD-10: F41.1). Covariate balance after propensity score matching was assessed using SMDs, with an absolute SMD <0.1 considered indicative of ideal balance and values <0.2 regarded as acceptable for residual imbalance. Follow-up duration for both cohorts was reported as the mean (± standard deviation) and the median (with interquartile range), expressed in days.
The primary outcome was the diagnosis of Dry Eye Disease (ICD10: H04.12). Statistical analyses were conducted using the TriNetX analytics platform. Hazard ratios (HRs) with 95% CIs were estimated using Cox proportional hazards regression models, and time-to-event outcomes were summarized with Kaplan–Meier cumulative incidence curves and compared using log-rank test. Differences in incidence proportions between cohorts were assessed with chi-square tests, and effect sizes were expressed as risk ratios (RRs) and odds ratios (ORs) with corresponding 95% confidence intervals (CIs). A 2-sided P value < .05 was considered statistically significant.
RESULTS
COHORT CHARACTERISTICS:
Cohort selection is summarized in Figure 1. Within the TriNetX US Collaborative Network, Cohort 1 included adult patients (≥18 years old) with a documented diagnosis of vitamin D deficiency (ICD-10: E55), totaling 6,634,934 individuals prior to propensity score matching. Cohort 2 included 11,069,982 adults with no history of vitamin D deficiency who had undergone routine adult medical examinations without abnormal findings (ICD-10: Z00.00). To align the study population with the 20-year observation window, 22,595 patients in Cohort 1 and 276,769 patients in Cohort 2 whose index event occurred before September 2005 were excluded. After propensity score matching, 6,047,502 patients were included in each cohort, yielding a combined matched study population of 12,095,004 individuals. To restrict analyses to incident DED (ICD-10: H04.12) and preserve temporal ordering between exposure and outcome, patients with a prior diagnosis of DED on or before the index date were excluded from the outcome analysis, resulting in the removal of 156,098 patients from Cohort 1 and 88,290 patients from Cohort 2. The final analytic sample, therefore, comprised 5,891,404 patients in the vitamin D–deficient cohort and 5,959,212 patients in the control cohort (total N =11,850,616).
FIGURE 1.
Cohort selection flowchart. The final analytic sample comprised 5,891,404 patients in the vitamin D–deficient cohort and 5,959,212 patients in the control cohort (total N = 11,850,616). Abbreviation: BMI: Body Mass Index.
Baseline characteristics were well balanced between the cohorts, with standardized mean differences (SMDs) <0.1 for most covariates; residual imbalances were observed for Black race (SMD = 0.122), BMI (SMD = 0.158), and not Hispanic or Latino ethnicity (SMD = 0.196), all of which remained below the acceptable 0.2 threshold, whereas unknown ethnicity showed a modest residual imbalance slightly above this cutoff (SMD = 0.238) (Table 1). In the post-match cohorts, the mean follow-up times were comparable, with 1,647 days (SD ±1,417) in the vitamin D deficiency group and 1,690 days (SD ±1,504) in controls. The corresponding median follow-up times were 1,261 days (IQR 2,059) and 1,294 days (IQR 2,201), respectively.
TABLE 1.
Pre and Post Matching Cohort Characteristics
| Characteristic | Vitamin D Deficiency (Pre-Match) N = 6,634,934 |
No Deficiency (Pre-Match) N = 11,069,982 |
SMD | Vitamin D Deficiency (Post-Match) N = 6,047,502 |
No Deficiency (Post-Match) N = 6,047,502 |
SMD |
|---|---|---|---|---|---|---|
| Age at Index (Mean ± SD) | 51.8 ± 19.9 | 43.8 ± 19.7 | 0.403 | 50.3 ± 19.7 | 50.8 ± 19.5 | 0.025 |
| Female (%) | 4,455,822 (67.4%) | 5,669,254 (52.5%) | 0.307 | 3,941,883 (65.2%) | 3,875,091 (64.1%) | 0.023 |
| Black or African American (%) | 1,033,092 (15.6%) | 1,265,073 (11.7%) | 0.114 | 952,227 (15.7%) | 698,462 (11.5%) | 0.122 |
| White (%) | 4,458,999 (67.4%) | 7,372,492 (68.3%) | 0.019 | 4,047,080 (66.9%) | 4,246,512 (70.2%) | 0.071 |
| American Indian or Alaska Native (%) | 31,661 (0.5%) | 36,074 (0.3%) | 0.023 | 29,268 (0.5%) | 19,248 (0.3%) | 0.026 |
| Native Hawaiian or Other Pacific | 19,026 (0.3%) | 36,567 (0.3%) | 0.009 | 17,380 (0.3%) | 20,185 (0.3%) | 0.008 |
| Islander (%) | ||||||
| Asian (%) | 297,824 (4.5%) | 462,698 (4.3%) | 0.011 | 277,524 (4.6%) | 245,936 (4.1%) | 0.026 |
| Other Race (%) | 265,861 (4.0%) | 449,365 (4.2%) | 0.007 | 249,356 (4.1%) | 229,076 (3.8%) | 0.017 |
| Unknown Race (%) | 505,876 (7.7%) | 1,170,944 (10.8%) | 0.111 | 474,667 (7.8%) | 588,083 (9.7%) | 0.066 |
| Not Hispanic or Latino (%) | 4,761,297 (72.0%) | 6,623,357 (61.4%) | 0.227 | 4,341,180 (71.8%) | 3,786,885 (62.6%) | 0.196 |
| Hispanic or Latino (%) | 541,872 (8.2%) | 860,643 (8.0%) | 0.008 | 507,095 (8.4%) | 441,629 (7.3%) | 0.040 |
| Unknown Ethnicity (%) | 1,309,170 (19.8%) | 3,309,213 (30.7%) | 0.252 | 1,199,227 (19.8%) | 1,818,988 (30.1%) | 0.238 |
| Menopausal and other | 25,274 (0.4%) | 15,888 (0.1%) | 0.046 | 19,314 (0.3%) | 15,316 (0.3%) | 0.012 |
| perimenopausal disorders (%) | ||||||
| Ischemic heart diseases (%) | 119,543 (1.8%) | 102,934 (1.0%) | 0.073 | 93,545 (1.5%) | 96,676 (1.6%) | 0.004 |
| Essential (primary) hypertension (%) | 558,985 (8.5%) | 438,827 (4.1%) | 0.182 | 443,001 (7.3%) | 376,709 (6.2%) | 0.044 |
| Diabetes mellitus (%) | 327,451 (5.0%) | 226,341 (2.1%) | 0.155 | 238,875 (3.9%) | 223,842 (3.7%) | 0.013 |
| Nicotine dependence (%) | 87,591 (1.3%) | 86,184 (0.8%) | 0.051 | 73,909 (1.2%) | 68,219 (1.1%) | 0.009 |
| Hypothyroidism (%) | 154,025 (2.3%) | 94,069 (0.9%) | 0.116 | 122,194 (2.0%) | 83,195 (1.4%) | 0.050 |
| Thyrotoxicosis [hyperthyroidism] (%) | 18,499 (0.3%) | 8,140 (0.1%) | 0.049 | 15,477 (0.3%) | 6,750 (0.1%) | 0.034 |
| Depressive episode (%) | 156,749 (2.4%) | 99,211 (0.9%) | 0.114 | 131,633 (2.2%) | 76,745 (1.3%) | 0.070 |
| Generalized anxiety disorder (%) | 59,706 (0.9%) | 45,444 (0.4%) | 0.059 | 52,364 (0.9%) | 30,653 (0.5%) | 0.043 |
| BMI | 30.5 ± 8.4 | 29.0 ± 7.3 | 0.183 | 30.4 ± 8.5 | 29.1 ± 7.3 | 0.158 |
SMD = standardized mean difference; BMI = body mass index; SD = standard deviation.
DRY EYE DISEASE INCIDENCE:
The incidence of DED was significantly higher in patients with vitamin D deficiency compared with matched controls (Figure 2). During follow-up, 196,639 of 5,891,404 patients (3.3%) in the vitamin D deficiency cohort and 160,141 of 5,959,212 patients (2.7%) in the control cohort developed DED. In time-to-event analysis using a Cox proportional hazards model, vitamin D deficiency was associated with an increased hazard of developing DED (HR 1.286, 95% CI 1.277–1.294; p < .001). Kaplan–Meier cumulative incidence of DED was likewise higher in the vitamin D deficiency cohort, with a significant log-rank test (χ2 = 5590.618; p < .001). The corresponding cumulative-incidence risk ratio and odds ratio at the end of follow-up were 1.242 (95% CI 1.234–1.250; p < .001) and 1.250 (95% CI 1.242–1.259; p < .001), respectively.
FIGURE 2.
Dry eye disease (DED) incidence in vitamin D deficiency vs controls. In the vitamin D deficiency cohort, 196,639 of 5,891,404 patients (3.3%) were diagnosed with dry eye disease (DED) following the index event, compared with 160,141 of 5,959,212 patients (2.7%) in the control cohort. This corresponded to a relative risk (RR) of 1.242 (95% confidence interval (CI): 1.234–1.250, P < .001) and an odds ratio (OR) of 1.250 (95% CI: 1.242–1.259, P < .001).
Cumulative incidence curves (Figure 3) demonstrated progressive risk accumulation of DED over time, with consistent separation between the vitamin D deficiency and control cohorts. At 1 year, cumulative incidence was 0.83% in the vitamin D deficiency cohort compared with 0.73% in controls. By 5 years, incidence rose to 3.58% versus 2.89%, and by 10 years reached 7.22% versus 5.52%. At 20 years (estimated via Kaplan–Meier methods), cumulative incidence was 17.87% in the vitamin D deficiency cohort compared with 14.16% in controls. Because relatively few patients contribute data at the far right tail of follow-up, these late-time estimates should be interpreted as extrapolated long-term trends under censoring rather than as fully observed proportions.
FIGURE 3.
Cumulative curve of incident dry eye disease (DED) diagnosis. Cumulative incidence curves demonstrated progressive risk accumulation of DED over time, with consistent separation between the vitamin D deficiency and control cohorts. At 1 year, cumulative incidence was 0.83% in the vitamin D deficiency cohort compared with 0.73% in controls. By 5 years, incidence rose to 3.58% versus 2.89%, and by 10 years reached 7.22% versus 5.52%. At 20 years of follow-up, cumulative incidence was 17.87% in the vitamin D deficiency cohort compared with 14.16% in controls, reflecting sustained divergence of the curves across long-term follow-up.
DISCUSSION
In this large, real-world cohort of 12,095,004 matched adults, vitamin D deficiency was associated with a higher incidence of DED in matched cohorts, with time-to-event analysis showing a 28.6% higher hazard (HR, 1.286, 95% CI: 1.277–1.294, p < .001). Although the association was modest, it was reproducible over extended follow-up. The data support an association between vitamin D deficiency and increased risk of DED in routine US healthcare.
Our findings align with and extend prior literature; prior work has described that both DED and MGD are common in the general population.2 Population-based studies show that reported DED prevalence varies widely world-wide, roughly from about 5% to 50%, and estimates tend to be somewhat higher in Asian populations than in other regions.17 Across studies stratified by sex, women consistently have a higher prevalence than men, and Asian ethnicity is identified as a largely consistent risk factor in global comparisons.17 In the US, a systematic review and meta-analysis reported a pooled DED prevalence of about 8.1% and an MGD prevalence of 21.2%, with 5-year DED incidence estimates ranging from 3.5% in adults ≥18 years to 7.8% in adults ≥68 years.2 Our study, a US-based cohort of adults aged ≥18 years, found an estimated 5-year Kaplan–Meier cumulative incidence of DED of 3.58% in the vitamin D–deficient cohort and 2.89% in controls, which is broadly comparable to prior 5-year US adult estimates, with a modestly higher incidence in the deficient cohort; the corresponding 20-year Kaplan–Meier estimates (17.87% vs 14.16%) are higher than short-term figures as the time horizon extends, but should be interpreted cautiously because they reflect extrapolated long-term trends under censoring, given substantially shorter observed median follow-up in this real-world Electronic Health Record (EHR) dataset.
Vitamin D has been implicated as a potential risk factor for DED, with small observational studies linking lower serum levels to worse symptoms, reduced tear production, and sometimes reduced tear stability, while supplementation studies suggest improvement in tear-film measures and discomfort in some patients.11,12,15,16,18,19 One randomized controlled trial enrolled 100 patients with DED and serum vitamin D deficiency (<20 ng/mL) and compared therapy with artificial tears alone to artificial tears plus oral vitamin D. Both groups used artificial tears 4 times daily, while the intervention group additionally received 50,000 IU of oral vitamin D once weekly for 8 weeks.18 After 8 weeks, the vitamin D group showed significantly greater improvement in several DED signs compared to controls: TBUT increased by 3.96 ± 1.48 seconds versus 0.70 ± 0.86 seconds in controls; Schirmer scores increased by 2.38 ± 1.55 mm versus 0.92 ± 1.57 mm; and tear osmolarity decreased by −16.9 ± 6.28 mOsm/L versus −3.34 ± 2.07 mOsm/L (p < .001 for all).18 Other trials have reported similar results. In a prospective randomized study of 90 patients with DED after Laser-Assisted In Situ Keratomileusis (LASIK), oral vitamin D3 2000 IU/day for 12 weeks significantly improved both symptoms and signs compared with no supplementation.19 At 3 months, mean OSDI scores were lower in the vitamin D than control group (10.25 ± 5.49 vs 20.22 ± 6.23), with higher TBUT (10.75 ± 1.09 vs 8.36 ± 1.23 s) and Schirmer values (11.34 ± 0.39 vs 8.12 ± 0.50 mm; all P < .05), and these differences persisted at 6 months.19 Such studies raise the possibility that vitamin D status may influence the ocular surface, but do not establish whether documented vitamin D deficiency is linked to incident, clinically diagnosed DED. In this context, our analysis uses diagnosis code–defined vitamin D deficiency and a clinical diagnosis of incident DED in a large, diverse US EHR network with predefined matching and long follow-up, reinforcing a link between vitamin D and incident DED, with an association that persists over time, even after adjustment for key comorbidities, findings that can be leveraged to support future therapeutic interventions.
There is biological plausibility for a link between vitamin D deficiency and DED, because vitamin D is locally metabolized in ocular tissues and the vitamin D receptor (VDR), the nuclear receptor that mediates calcitriol’s gene-regulating, anti-inflammatory effects, is expressed in the eye.9 Experimental studies indicate that active vitamin D (calcitriol) can dampen hyperosmotic-stress–induced corneal epithelial inflammation and oxidative stress and reduce epithelial apoptosis via VDR-dependent pathways, supporting a role for vitamin D signaling in maintaining ocular surface inflammatory homeostasis.9 Work on nutrition and the ocular surface further suggests that micronutrient balance, including vitamin D status, may influence ocular surface ecology and the ocular microbiome.10 Taken together, these basic science and preclinical data support a mechanistic link between vitamin D pathways and ocular surface homeostasis, providing biologic plausibility for the association observed in our cohort.
As with all studies, our findings should be interpreted considering both our limitations and strengths. Limitations include potential exposure misclassification (ICD-10–based deficiency rather than measured 25-hydroxyvitamin D levels), outcome misclassification (diagnosis codes rather than standardized signs/symptoms or subtype), and residual confounding (e.g., autoimmune activity, medications such as anticholinergics or isotretinoin, environmental factors, screen time, and contact lens wear). Furthermore, ICD-10 based coding does not allow us to further examine which aspect of DED (symptoms, stability, or tear production) is most impacted by Vitamin D deficiency. Although baseline characteristics were generally well balanced after matching (SMDs <0.1 for most covariates), Black race, BMI, and not Hispanic/Latino ethnicity showed small residual imbalances (all SMDs <0.2 within the acceptable range), and unknown ethnicity had a modest imbalance slightly above this threshold (SMD =0.238), so some residual confounding from these factors cannot be excluded. Differential health-system contact among vitamin D-deficient patients could also introduce ascertainment bias. The study also lacked reliable data on supplementation or repletion, which may have biased the estimates. As a retrospective EHR study, causality cannot be inferred, and site-level coding or practice heterogeneity may reduce the generalizability of these findings. Furthermore, social determinants of health, which are key drivers of disease incidence and progression, are not captured within the TriNetX database. However, our study also had many strengths including a large scale and representativeness population of US adults engaged in care across participating health systems, predefined propensity matching with balance assessment, and a focus on incident DED that clarifies temporal ordering.
This study demonstrates a significant association between vitamin D deficiency and an increased risk of DED incidence, highlighting the potential role of vitamin D supplementation in managing and preventing aspects of DED in appropriate populations. Given the widespread prevalence of vitamin D deficiency globally, early screening and intervention strategies may improve ocular surface health and prevent vision-related discomfort associated with DED, particularly in at-risk groups (limited sun exposure, higher latitude, darker skin pigmentation, malabsorption, obesity, or autoimmune disease).20 In combination with calcium, vitamin D supplementation has been shown to improve other aspects of health, such as increasing bone mineral density (BMD) in older adults.20 By analogy, in patients with DED who are at risk for deficiency, identifying and correcting low vitamin D levels may be a reasonable adjunct to standard DED therapies, while recognizing that supplementation should be guided by general medical indications rather than used as a stand-alone treatment for DED. However, future research is necessary to explore causal mechanisms, optimize supplementation protocols, examine dose–response relationships using measured 25-hydroxyvitamin D levels, and evaluate long-term outcomes of vitamin D intervention in DED. Additional studies could also explore the role of vitamin D in varying degrees of DED severity and sub-types (e.g., autoimmune associated, post-menopausal status) and its interaction with other medications. Addressing micronutrient imbalances through targeted therapies and public health initiatives could be crucial for reducing the incidence and severity of DED, enhancing patient quality of life.
CONCLUSION
This large, nationwide retrospective cohort study demonstrates a significant association between vitamin D deficiency and an elevated risk of incident dry eye disease. The consistently higher cumulative incidence and 28.6% increased hazard of DED in vitamin D–deficient adults, reproduced over long-term follow-up in a matched U.S. population, underscores vitamin D deficiency as a potential independent and modifiable risk factor for ocular surface disease. These findings reinforce the clinical relevance of systemic nutritional status in ocular surface homeostasis and suggest that identifying and correcting vitamin D deficiency may be a reasonable adjunct to standard DED prevention and management in at-risk patients. Future prospective studies and randomized supplementation trials are needed to confirm causality, define dose–response relationships using measured 25-hydroxyvitamin D levels, and determine whether vitamin D repletion can meaningfully reduce DED incidence or severity, ultimately informing targeted strategies to preserve ocular comfort, visual function, and quality of life.
Supplementary Material
Supplemental Material available at AJO.com.
Acknowledgment:
We confirm that all coauthors have reviewed the revised version of this manuscript, have seen and agree with all changes made in this revision, and approve the manner in which their names and authorship positions are listed.
Funding/Support:
This research was funded in part by NIH Center Core Grant (P30EY013810); Research to Prevent Blindness Unrestricted Grant.
Footnotes
DECLARATION OF COMPETING INTEREST
The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.
CREDIT AUTHORSHIP CONTRIBUTION STATEMENT
Shon B. Shmushkevich: Writing – original draft, Visualization, Validation, Supervision, Software, Resources, Project administration, Methodology, Investigation, Funding acquisition, Formal analysis, Data curation, Conceptualization. Sandi Bajrami: Writing – review & editing, Writing – original draft, Visualization, Validation, Supervision, Software, Resources, Project administration, Methodology, Investigation, Formal analysis, Data curation, Conceptualization. Benjamin Beauchamp: Writing – review & editing, Writing – original draft, Software, Project administration. Julia S. Kane: Writing – review & editing, Writing – original draft, Supervision, Software. Dima L. Mutawe: Writing – review & editing, Writing – original draft, Software. James C. Hollinger: Writing – review & editing, Writing – original draft. Victoria A. Marks: Writing – review & editing, Writing – original draft, Conceptualization. Prashant K. Parekh: Writing – review & editing, Writing – original draft, Funding acquisition, Formal analysis, Data curation, Conceptualization. Adam L. Rothman: Writing – review & editing, Writing – original draft, Project administration. Zubair A. Ansari: Writing – review & editing. Basil K. Williams Jr: Writing – review & editing, Writing – original draft, Visualization, Validation, Supervision, Software, Resources, Project administration. Anat Galor: Writing – review & editing, Writing – original draft, Visualization, Validation, Supervision, Software, Resources, Project administration, Methodology, Investigation, Funding acquisition, Formal analysis, Data curation, Conceptualization.
Financial Disclosures: The authors indicate no financial support or conflicts of interest. All authors attest that they meet the current ICMJE criteria for authorship.
REFERENCES
- 1.Craig JP, Nichols KK, Akpek EK, et al. TFOS DEWS II definition and classification report. Ocul Surf. 2017;15(3):276–283. [DOI] [PubMed] [Google Scholar]
- 2.McCann P, Abraham AG, Mukhopadhyay A, et al. Prevalence and incidence of dry eye and meibomian gland dysfunction in the united states: a systematic review and meta-analysis. JAMA Ophthalmol. 2022;140(12):1181–1192. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Galor A, Levitt RC, Felix ER, Sarantopoulos CD. Understanding the true burden of dry eye disease. Expert Rev Ophthalmol. 2015;10(5):403–405. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Yu J, Asche CV, Fairchild CJ. The economic burden of dry eye disease in the United States: a decision tree analysis. Cornea. 2011;30(4):379–387. [DOI] [PubMed] [Google Scholar]
- 5.Pouyeh B, Viteri E, Feuer W, et al. Impact of ocular surface symptoms on quality of life in a United States Veterans Affairs population. Am J Ophthalmol. 2012;153(6):1061–1066 e1063. [DOI] [PubMed] [Google Scholar]
- 6.Schiffman RM, Walt JG, Jacobsen G, Doyle JJ, Lebovics G, Sumner W. Utility assessment among patients with dry eye disease. Ophthalmology. 2003;110(7):1412–1419. [DOI] [PubMed] [Google Scholar]
- 7.Jones L, Downie LE, Korb D, et al. TFOS DEWS II management and therapy report. Ocul Surf. 2017;15(3):575–628. [DOI] [PubMed] [Google Scholar]
- 8.McCann P, Kruoch Z, Lopez S, Malli S, Qureshi R, Li T. Interventions for dry eye: an overview of systematic reviews. JAMA Ophthalmol. 2024;142(1):58–74. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Caban M, Lewandowska U. Vitamin D, the Vitamin D receptor, calcitriol analogues and their link with ocular diseases. Nutrients. 2022;14(11):2353. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Pilkington M, Lloyd D, Guo B, Watson SL, Ooi KG-J. Effects of dietary imbalances of micro- and macronutrients on the ocular microbiome and its implications in dry eye disease. Exploration of Medicine. 2024;5(1):127–147. [Google Scholar]
- 11.Kurtul BE, Özer PA, Aydinli MS. The association of vitamin D deficiency with tear break-up time and Schirmer testing in non-Sjögren dry eye. Eye (Lond). 2015;29(8):1081–1084. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Liu J, Dong Y, Wang Y. Vitamin D deficiency is associated with dry eye syndrome: a systematic review and meta-analysis. Acta Ophthalmol. 2020;98(8):749–754. [DOI] [PubMed] [Google Scholar]
- 13.Jeon DH, Yeom H, Yang J, Song JS, Lee HK, Kim HC. Are serum vitamin D levels associated with dry eye disease? Results from the Study Group for Environmental Eye Disease. J Prev Med Public Health. 2017;50(6):369–376. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Galor A, Gardener H, Pouyeh B, Feuer W, Florez H. Effect of a Mediterranean dietary pattern and vitamin D levels on dry eye syndrome. Cornea. 2014;33(5):437–441. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Chen Z, Zhang C, Jiang J, et al. The efficacy of vitamin D supplementation in dry eye disease: a systematic review and meta-analysis. Cont Lens Anterior Eye. 2024;47(5):102169. [DOI] [PubMed] [Google Scholar]
- 16.Bae SH, Shin YJ, Kim HK, Hyon JY, Wee WR, Park SG. Vitamin D supplementation for patients with dry eye syndrome refractory to conventional treatment. Sci Rep. 2016;6:33083. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17.Stapleton F, Alves M, Bunya VY, et al. TFOS DEWS II Epidemiology Report. Ocul Surf. 2017;15(3):334–365. [DOI] [PubMed] [Google Scholar]
- 18.Najjaran M, Zarei-Ghanavati S, Arjmand Askari E, Eslampoor A, Ziaei M. Effect of oral vitamin D supplementation on dry eye disease patients with vitamin D deficiency. Clin Exp Optom. 2023;106(3):257–262. [DOI] [PubMed] [Google Scholar]
- 19.Lin Y, Su H, Wu J, Yuan M, Zhang Y. Oral vitamin D(3) supplementation for femtosecond LASIK-associated dry eye vitamin D for LASIK dry eye syndrome. Int Ophthalmol. 2022;42(10):3145–3152. [DOI] [PubMed] [Google Scholar]
- 20.Chung M, Balk EM, Brendel M, et al. Vitamin D and calcium: a systematic review of health outcomes. Evid Rep Technol Assess (Full Rep). 2009(183):1–420. [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.