Key Points
Question
Is there an association between hypovitaminosis D and noninfectious uveitis in a large national sample?
Findings
In this case-control study of 558 cases of noninfectious uveitis and 2790 controls, having an adequate vitamin D level was associated with 21% lower odds of having a case of noninfectious uveitis.
Meaning
Hypovitaminosis D may be a risk factor for noninfectious uveitis.
Abstract
Importance
Understanding the role of vitamin D—which regulates inflammatory responses—in noninfectious uveitis (an inflammatory disease) may provide insight into treatment and prevention of this disease.
Objective
To investigate whether there is an association between hypovitaminosis D and incident noninfectious uveitis.
Design, Setting, and Participants
In a retrospective case-control study, data from a health care claims database containing deidentified medical claims from a large private insurer were used to identify 558 adults enrolled from January 1, 2000, to December 31, 2016, who received a diagnosis of noninfectious uveitis from an eye care clinician (with receipt of a confirmatory diagnosis within 120 days of the initial diagnosis) and who had a vitamin D level measured within 1 year before the first diagnosis. Exclusion criteria included having systemic disease or receiving medication known to lower vitamin D levels, having undergone intraocular surgery, and having infectious uveitis. Each case patient was matched with 5 controls on the basis of age, sex, race/ethnicity, and index date (2790 controls). The controls had vitamin D level determined either within 1 year before or within 6 months after receiving an eye examination with normal findings. Multiple logistic regression models were used to examine the association between hypovitaminosis D and noninfectious uveitis.
Main Outcomes and Measures
The primary, prespecified analysis assessed the association of noninfectious uveitis with hypovitaminosis D (vitamin D level ≤20 ng/mL).
Results
The 558 cases and 2790 controls were matched on age, and each group had a mean (SD) age of 58.9 (14.7) years. Among the cohort of 3348 patients, 2526 (75.4%) were female, and the racial/ethnic distribution in the matched samples was 2022 (60.4%) white, 552 (16.5%) black, 402 (12.0%) Hispanic, 162 (4.8%) Asian, and 210 (6.3%) unknown. Patients with normal vitamin D levels had 21% lower odds of having noninfectious uveitis than patients with low vitamin D levels (odds ratio [OR], 0.79; 95% CI, 0.62-0.99; P = .04). In a race-stratified analysis, an association between vitamin D and uveitis was found in black patients (OR, 0.49; 95% CI, 0.30-0.80; P = .004) and was qualitatively similar but nonsignificant in white patients (OR, 0.87; 95% CI, 0.62-1.21; P = .40) and Hispanic patients (OR, 0.60; 95% CI, 0.33-1.10; P = .10).
Conclusions and Relevance
This and other reports have found an association between hypovitaminosis D and noninfectious uveitis. However, these studies cannot establish a causal relationship. Prospective studies are warranted to evaluate whether hypovitaminosis D causes increased risk of uveitis and the role of vitamin D supplementation in prevention and treatment of uveitis.
This case-control study uses a large health care claims database to investigate whether there is an association between hypovitaminosis D and an increased risk of uveitis in adults who received a diagnosis of noninfectious uveitis.
Introduction
Noninfectious uveitis is largely a T-cell–mediated disease.1 The vitamin D receptor is present on most types of immune cells, including T lymphocytes.2,3 In mouse studies, treatment with vitamin D prevented or mitigated uveitis.4 A previous study demonstrated an association between hypovitaminosis D and the presence of noninfectious anterior uveitis in a relatively small sample at a single geographic location.5 To confirm this observation, we conducted a case-control study in a large population from across the United States.
Methods
Our study involved patients enrolled from January 1, 2000, to December 31, 2016, in the Clinformatics Data Mart Database (OptumInsight), which contains the deidentified claims of beneficiaries from a large private insurer. The institutional review boards of the Massachusetts Eye and Ear Infirmary and the University of Pennsylvania deemed this study exempt from review and waived the need for patient informed consent because the data were deidentified.
In this retrospective case-control study, case patients were 18 years or older and had received a diagnosis of uveitis from an eye care clinician followed by a second uveitis diagnosis within 120 days of the initial diagnosis (to confirm that the diagnosis was not made simply to rule out uveitis; see Table 1 for all diagnosis codes used in this study). The date of the first diagnosis was the index date. Cases were also required to have at least 1 vitamin D level determined within 1 year before the index date; the laboratory blood sample obtained closest to the index date was used for analysis. Potential cases were excluded for having less than 24 months of data in the database before the index date, having a systemic disease, or receiving medication (phenobarbital, orlistat, phenytoin, or corticosteroids) known to decrease vitamin D levels. To further reduce miscategorization of noninfectious uveitis, patients who had intraocular surgery within the 90 days before the index date or who received a diagnosis of an infectious systemic disease or infectious uveitis at any time were also excluded (Table 1).
Table 1. ICD-9 and ICD-10 Codes Used in This Study.
| Diagnosis | ICD-9 Codes | ICD-10 Codes |
|---|---|---|
| Noninfectious uveitis (case inclusion diagnoses) | ||
| Anterior uveitis | 363.22, 364.0x, 364.1x, 364.20-2, 364.25-9, 364.3x, 364.4x | H20.00-2, H20.04-9, H20.1x, H20.80-1, H20.83-9, H20.9x, H30.81 |
| Intermediate uveitis | 363.21 | H30.2 |
| Posterior uveitis | 360.11, 360.12, 360.18, 363.20, 364.24 | H20.82, H30.80, H30.82-9, H30.9x, H35.06, H44.13 |
| Infectious and trauma-related diseases (case exclusion diagnoses) | ||
| Herpetic eye disease | 053.2x, 053.4x | B02.3x |
| Syphilis | 09x.xx, 104.00, 647.0x | A50.xx, A51.xx, A52.xx |
| Infectious iridiocyclitis | 098.41, 364.03, 364.23 | A54.32, H20.2x, H20.39 |
| Histoplasmosis | 115.92 | B39.9x, H32.xx |
| Chorioretinitis (including toxoplasma chorioretinitis) | 130.2x, 363.0x, 363.10-3 | B58.01, H33.0x, H33.1x, P37.1x, |
| Endophthalmitis | 360.19 | H20.03 |
| CMV retinitis | 078.5 | B25.9x |
| Trauma related | 921.3x | S05.xx |
| Tuberculosis | 01x.xx, 137.xx, 647.3x | A15.xx, A17.xx, A18.xx, A19.xx |
| Lyme disease | 088.81 | A69.2x |
| Systemic diagnoses identified as associated with low 25-hydroxy-vitamin D levels | ||
| Parathyroid dysfunction | 252.xx | E21.xx |
| Sarcoidosis | 135.00, 321.40 | D86.x |
| Renal failure | 285.21, 584.xx, 585.6x, 585.9x | N17.xx, N18.6x, N19.xx |
| Gastric bypass | V45.3x, V45.86 | Z98.xx |
| Systemic diagnoses identified as associated with high 25-hydroxy-vitamin D levels | ||
| Rheumatoid arthritis | 714.xx | M05.xx |
| Behçet disease | 711.2x, 136.10 | M35.2x |
| Juvenile idiopathic arthritis | 714.xx | M08.xx |
| Reactive arthritis (Reiter disease) | 099.30, 711.1x | M02.xx |
| Ankylosing spondylitis | 720.00 | M45.xx |
| Inflammatory bowel disease | 555.xx, 556.xx | K50.xx, K51.xx, K52.xx |
| Multiple sclerosis | 340.00 | G35.xx |
| Systemic lupus erythematosus | 373.34, 695.40, 710.00 | M32.xx, L93.0x |
| Granulomatosis with polyangiitis | 446.40 | M31.3x |
| Relapsing polychondritis | 733.99 | M94.1x |
| Tubulointerstitial nephritis | 583.89, 583.9x | N12.xx |
| Whipple disease | 040.20 | M14.8x |
Abbreviations: CMV, cytomegalovirus; ICD-9, International Classification of Diseases, Ninth Revision; ICD-10, International Statistical Classification of Diseases and Related Health Problems, Tenth Revision.
Five controls were matched to each case using age (±3 years), sex, and race/ethnicity because these demographic factors have been associated with both vitamin D levels and uveitis.6,7,8,9,10 A 5:1 ratio was chosen because power did not increase substantially for higher ratios. Controls were required to have been examined by an eye care clinician before the index date of the matched case. The same exclusion criteria were applied to cases and controls, except that controls were allowed to have a vitamin D level determined up to 6 months after the index date of the matched case (instead of only before that date). This was done because vitamin D was not a predictive factor of disease for controls, and we wanted to maximize the power of our statistical analysis.
The primary outcome assessed the association of noninfectious uveitis with hypovitaminosis D. Vitamin D was evaluated as a dichotomous variable, with a 25-hydroxy vitamin D level of 20 ng/mL (to convert to nanomoles per liter, multiply by 2.496) or greater to be considered adequate because that level of vitamin D is sufficient for bone and overall health.11 Conditional logistic regression, adjusted for 2 proxies for sun exposure, was used to assess the association between uveitis and hypovitaminosis D with SAS, version 9.4 (SAS Institute Inc). The 2 proxies of sun exposure were geographic region and month of laboratory blood draw: fewer hours of daylight (November-March) vs regular hours of daylight (April-October). Because of deidentification concerns, the database provides information on geographic regions representing groups of states. The eastern states are grouped into northern and southern blocks, which is optimal for this analysis of vitamin D, but the western states are not divided into north and south. Therefore, for our analyses, we divided the country into 3 regions: Northeast, Southeast and West. In addition, we conducted race-stratified analyses and a sensitivity analysis that excluded patients with systemic diseases that are thought to increase vitamin D levels.
Results
Five hundred fifty-eight patients with uveitis were matched with 2790 controls. Among the combined cohort of 3348 patients, who had a mean (SD) age of 58.9 (14.7) years, 2526 (75.4%) were female, and the racial/ethnic distribution in the matched samples was 2022 (60.4%) white, 552 (16.5%) black, 402 (12.0%) Hispanic, 162 (4.8%) Asian, and 210 (6.3%) unknown. Compared with patients with noninfectious uveitis who did not have a vitamin D level determined within 1 year before diagnosis, patients included in the study were older (58.9 vs 52.4 years), more likely to be female (75.4% vs 54.1%), and more likely to be nonwhite (39.6% vs 33.1%) (P < .001 for all comparisons). No imbalances were seen between cases and matched controls with respect to demographic characteristics (Table 2). Table 3 presents the association results. Overall, patients with an adequate vitamin D level had 21% lower odds of having uveitis compared with patients with low vitamin D levels (odds ratio [OR], 0.79; 95% CI, 0.62-0.99; P = .04). In the race-stratified analysis, the association between low vitamin D levels and having uveitis was significant among black patients (OR, 0.49; 95% CI, 0.30-0.80; P = .004) and was nonsignificant but qualitatively similar in white patients (OR, 0.87; 95% CI, 0.62-1.21; P = .40) and Hispanic patients (OR, 0.60; 95% CI, 0.33-1.10; P = .10). In the sensitivity analysis of the 481 cases and 2058 controls without any systemic diseases thought to increase vitamin D levels, patients with an adequate vitamin D level had 25% lower odds of having uveitis (OR, 0.75; 95% CI, 0.58-0.97; P = .03).
Table 2. Study Patient Characteristics.
| Variable | No. (%) of Patients | |
|---|---|---|
| Controls (n = 2790) | Cases (n = 558) | |
| Age, y | ||
| <50 | 793 (28.4) | 159 (28.5) |
| 50-64 | 967 (34.7) | 193 (34.6) |
| ≥65 | 1030 (37.0) | 206 (36.9) |
| Sex | ||
| Male | 685 (24.6) | 137 (24.6) |
| Female | 2105 (75.4) | 421 (75.4) |
| Race/ethnicity | ||
| White | 1685 (60.4) | 337 (60.4) |
| Black | 460 (16.5) | 92 (16.5) |
| Hispanic | 335 (12.0) | 67 (12.0) |
| Asian | 135 (4.8) | 27 (4.8) |
| Unknown | 175 (6.3) | 35 (6.3) |
| US geographic region | ||
| Northeast | 611 (21.9) | 102 (18.3) |
| Southeast | 1565 (56.1) | 338 (60.6) |
| West | 601 (21.5) | 116 (20.8) |
| Unknown | 13 (0.5) | 2 (0.4) |
| Vitamin D blood draw month | ||
| April-October | 1563 (56.0) | 301 (53.9) |
| November-March | 1227 (44.0) | 257 (46.1) |
Table 3. Unadjusted and Adjusted Logistic Regression Results.
| Variable | No. of Patients | Odds Ratio (95% CI) | P Value |
|---|---|---|---|
| Unadjusted logistic regression | |||
| Vitamin D level | |||
| <20 ng/dL | 587 | 1 [Reference] | .03 |
| ≥20 ng/dL | 2761 | 0.78 (0.62-0.98) | |
| Adjusted logistic regressiona | |||
| Vitamin D level | |||
| <20 ng/dL | 587 | 1 [Reference] | .04 |
| ≥20 ng/dL | 2761 | 0.79 (0.62-0.99) | |
| Vitamin D blood draw month | |||
| April-October | 1864 | 1 [Reference] | .39 |
| November-March | 1484 | 1.08 (0.90-1.30) | |
| Geographic region | |||
| Northeast | 713 | 1 [Reference] | .18 |
| Southeast | 1903 | 1.31 (1.02-1.67) | |
| West | 717 | 1.16 (0.86-1.56) | |
| Unknown | 9 | 0.91 (0.20-4.11) | |
| Adjusted race/ethnicity–stratified logistic regressiona,b | |||
| Vitamin D level in white patients | |||
| <20 ng/dL | 265 | 1 [Reference] | .40 |
| ≥20 ng/dL | 1757 | 0.87 (0.62-1.47) | |
| Vitamin D level in black patients | |||
| <20 ng/dL | 150 | 1 [Reference] | .004 |
| ≥20 ng/dL | 402 | 0.49 (0.30-0.80) | |
| Vitamin D level in Hispanic patients | |||
| <20 ng/dL | 51 | 1 [Reference] | .10 |
| ≥20 ng/dL | 111 | 0.60 (0.33-1.10) |
SI conversion factor: To convert vitamin D to nanomoles per liter, multiply by 2.496.
Adjusted for geographic region and month of laboratory blood draw.
Asian and unknown categories were not run owing to small sample sizes.
Discussion
In this study, noninfectious uveitis was found to be associated with hypovitaminosis D. Because cases had their vitamin D level determined within 1 year before the index date, this association might indicate a causal role of low levels of vitamin D in the development of uveitis. A number of immunomodulatory mechanisms could explain these findings, including the ability of vitamin D to upregulate anti-inflammatory type 2 helper T cells and suppress inflammatory cytokine activity.
Smaller investigations have shown associations between hypovitaminosis D and specific uveitis subtypes.5,12,13,14 This larger study confirms earlier findings. It also provides some suggestion that race/ethnicity might affect the association between vitamin D level and noninfectious uveitis. When different racial/ethnic factors were evaluated, we found that the association between low vitamin D levels and noninfectious uveitis was statistically significant only for black patients. This finding could indicate a stronger biological effect of vitamin D on uveitis in this race, or it might indicate limited power to detect a similar difference in the other racial/ethnic groups, especially given that an earlier report5 found the same association in a population that was mostly white. The effect of race on the association between low vitamin D levels and noninfectious uveitis is an important area for future investigation.
Limitations
Limitations of our study include our inability to assess supplemental use of vitamin D. Although vitamin D levels were determined more than once, we did not believe that these data would be a reliable surrogate for assessing vitamin D supplementation. Some patients with hypovitaminosis D might have started receiving vitamin D supplementation shortly after vitamin D levels were determined, which might have changed vitamin D levels by the index date and resulted in a miscategorization of such patients as having low levels of vitamin D. However, such a bias would be toward the null and would suggest that our results may underestimate the true association between hypovitaminosis D and uveitis. The retrospective case-control design of our study, compared with prospective designs, is less reliable for establishing causality. Use of an extant data set prevented our study from accounting for smoking history (which was unavailable in the data set) and prevented confirmation of the uveitis diagnosis by medical record review. Only 3% to 5% of patients in this database had vitamin D levels determined. Patients typically have vitamin D levels determined as part of preventive primary care or because of a medical indication, such as metabolic bone disease or nutrient absorption problems. While patients with data on vitamin D level are therefore likely to differ from the general population by being more compliant with primary care or having one of the conditions listed above, such differences should affect cases and controls equally, which would also affect the generalizability of our findings but not their validity.
Conclusions
This large national study suggests that hypovitaminosis D might increase the risk of uveitis. Prospective studies are needed to better evaluate causality. If causality were to be established, vitamin D supplementation would be an attractive potential adjunct intervention in the treatment of noninfectious uveitis.
References
- 1.Nussenblatt RB. Basic and clinical immunology in uveitis. Jpn J Ophthalmol. 1987;31(3):368-374. [PubMed] [Google Scholar]
- 2.Deluca HF, Cantorna MT. Vitamin D: its role and uses in immunology. FASEB J. 2001;15(14):2579-2585. [DOI] [PubMed] [Google Scholar]
- 3.Veldman CM, Cantorna MT, DeLuca HF. Expression of 1,25-dihydroxyvitamin D(3) receptor in the immune system. Arch Biochem Biophys. 2000;374(2):334-338. [DOI] [PubMed] [Google Scholar]
- 4.Tang J, Zhou R, Luger D, et al. . Calcitriol suppresses antiretinal autoimmunity through inhibitory effects on the Th17 effector response. J Immunol. 2009;182(8):4624-4632. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Grotting LA, Davoudi S, Palenzuela D, Papaliodis GN, Sobrin L. Association of low vitamin D levels with noninfectious anterior uveitis. JAMA Ophthalmol. 2017;135(2):150-153. [DOI] [PubMed] [Google Scholar]
- 6.Tsirouki T, Dastiridou A, Symeonidis C, et al. . A focus on the epidemiology of uveitis. Ocul Immunol Inflamm. 2018;26(1):2-16. [DOI] [PubMed] [Google Scholar]
- 7.Ishaq M, Muhammad JS, Mahmood K. Uveitis is not just an ophthalmologists’ concern. J Pak Med Assoc. 2012;62(2):92-97. [PubMed] [Google Scholar]
- 8.Rathinam SR, Namperumalsamy P. Global variation and pattern changes in epidemiology of uveitis. Indian J Ophthalmol. 2007;55(3):173-183. [DOI] [PubMed] [Google Scholar]
- 9.Jiménez-Balderas FJ, Fernandez-Arrieta G, Camargo-Coronel A, et al. . Uveitis in adult patients with rheumatic inflammatory autoimmune diseases at a tertiary-care hospital in Mexico City. J Rheumatol. 2011;38(2):325-330. [DOI] [PubMed] [Google Scholar]
- 10.Päivönsalo-Hietanen T, Tuominen J, Saari KM. Seasonal variation of endogenous uveitis in south-western Finland. Acta Ophthalmol Scand. 1998;76(5):599-602. [DOI] [PubMed] [Google Scholar]
- 11.Ross AC, Manson JE, Abrams SA, et al. . The 2011 report on dietary reference intakes for calcium and vitamin D from the Institute of Medicine: what clinicians need to know. J Clin Endocrinol Metab. 2011;96(1):53-58. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Dadaci Z, Cetinkaya S, Oncel Acir N, Oncel M, Borazan M. Serum vitamin D levels in patients with acute anterior uveitis. Ocul Immunol Inflamm. 2017;25(4):492-496. [DOI] [PubMed] [Google Scholar]
- 13.Mitulescu TC, Stavaru C, Voinea LM, Banica LM, Matache C, Predeteanu D. The role of Vitamin D in immuno-inflammatory responses in ankylosing spondylitis patients with and without acute anterior uveitis. J Med Life. 2016;9(1):26-33. [PMC free article] [PubMed] [Google Scholar]
- 14.Yi X, Yang P, Sun M, Yang Y, Li F. Decreased 1,25-dihydroxyvitamin D3 level is involved in the pathogenesis of Vogt-Koyanagi-Harada (VKH) disease. Mol Vis. 2011;17:673-679. [PMC free article] [PubMed] [Google Scholar]