JIA indicates juvenile idiopathic arthritis.
aPatient sex was recorded by the insurance company.
eTable 1. ICD-9-CM Codes for Uveitis (Infectious and Noninfectious)
eTable 2. ICD-9-CM Codes for Noninfectious Uveitis
eTable 3. ICD-9-CM Codes Used to Identify Systemic Conditions
Customize your JAMA Network experience by selecting one or more topics from the list below.
Thorne JE, Suhler E, Skup M, et al. Prevalence of Noninfectious Uveitis in the United States: A Claims-Based Analysis. JAMA Ophthalmol. 2016;134(11):1237–1245. doi:10.1001/jamaophthalmol.2016.3229
What is the prevalence of noninfectious uveitis among adults and children stratified by age, sex, and location of inflammation?
This study used a large US administrative insurance claims database and found that the prevalence of noninfectious uveitis among adults was 121 cases per 100 000 persons and among children was 29 per 100 000 persons. Approximately 10% of anterior cases were found to require therapy above topical corticosteroids.
A better understanding of the prevalence of noninfectious uveitis will help to determine the number of patients affected by different types of this disease.
Noninfectious uveitis (NIU) is a collection of intraocular inflammatory disorders that may be associated with significant visual impairment. To our knowledge, few studies have investigated NIU prevalence overall or stratified by inflammation location, severity, presence of systemic conditions, age, or sex.
To estimate NIU prevalence using a large, retrospective, administrative claims database.
Design, Setting, and Participants
This analysis used the OptumHealth Reporting and Insights database to estimate 2012 NIU prevalence. Analysis was conducted in September 2016. The large administrative insurance claims database includes 14 million privately insured individuals in 69 self-insured companies spanning diverse industries. Included in the study were patients with NIU with 2 or more uveitis diagnoses on separate days in 2012 and continuous enrollment in a health plan for all of 2012 and categorized by inflammation site.
Main Outcomes and Measures
We estimated overall NIU prevalence by inflammation site, severity, sex, and age. Patients with anterior NIU were categorized by the presence of systemic conditions.
Of the approximately 4 million eligible adult patients, approximately 2.1 million were women, and of the 932 260 children, 475 481 were boys. The adult prevalence of NIU was 121 cases per 100 000 persons (95% CI, 117.5-124.3). The pediatric NIU prevalence was 29 cases per 100 000 (95% CI, 26.1-33.2). Anterior NIU accounted for 81% (3904 cases) of adult NIU cases (98 per 100 000; 95% CI, 94.7-100.9) and 75% (207 cases) of pediatric NIU cases (22 per 100 000; 95% CI, 19.3-25.4). The prevalences of noninfectious intermediate, posterior, and panuveitis were, for adults, 1 (95% CI, 0.8-1.5), 10 (95% CI, 9.4-11.5), and 12 (95% CI, 10.6-12.7) per 100 000, respectively, and for pediatric patients, 0 (95% CI, 0.1-1.1), 3 (95% CI, 1.8-4.1), and 4 (95% CI, 2.9-5.6) per 100 000, respectively. The prevalence of NIU increased with age and was higher among adult females than males. Application of these estimates to the US population suggests that NIU affected approximately 298 801 American adults (95% CI, 290 512-307 324) and 21 879 children (95% CI, 19 360-24 626) in 2015.
Conclusions and Relevance
The estimated prevalence of NIU was 121 cases per 100 000 for adults (95% CI, 117.5-124.3) and 29 per 100 000 for children (95% CI, 26.1-33.2). Prevalence was estimated using administrative claims from a commercially insured population, which may have a different prevalence than other segments of the US population. A better understanding of the prevalence of NIU will help to determine the number of patients affected.
Uveitis is a group of ocular disorders characterized by inflammation inside the eye, which can cause significant visual impairment.1 It is estimated to be responsible for 10% of cases of blindness in the United States,2,3 including 30 000 new cases of legal blindness each year.4 Uveitis is categorized based on the primary anatomical location of inflammation,5 with anterior uveitis accounting for 50% to 90% of all cases.6-8
Although uveitis may be infectious or noninfectious, noninfectious uveitis (NIU) is the most common type observed in the United States. Despite this, the prevalence of NIU has not been studied separately from infectious uveitis. Noninfectious uveitis is of particular interest because of the need for improved treatment options for many patients. Although patients with anterior NIU typically are responsive to topical corticosteroids, this treatment is not effective for patients with intermediate, posterior, or pan-NIU,9 and long-term corticosteroid use may be associated with potentially serious adverse effects.10 Furthermore, some patients with anterior NIU may require long-term suppressive doses of topical corticosteroids or therapies over and above topical corticosteroids to control intraocular inflammation and prevent structural ocular complications. Patients with long-term anterior, intermediate, posterior, and pan-NIU are also at risk for complications or vision loss. Thus, this study aimed to estimate the prevalence of NIU in the United States.
While there are a few published estimates of uveitis prevalence in the United States, to our knowledge, none focused specifically on NIU. In addition, most studies focused on particular US regions,2,11-13 while one, which used Medicare claims data, only considered patients age 65 years and older.14 The current study used a large database of medical and pharmacy claims to estimate the prevalence of NIU overall by location of inflammation, sex, and age.
This study used data from the OptumHealth Reporting and Insights database, an administrative database covering 14 million privately insured individuals in 69 self-insured companies, covering a range of industries and US regions. The database includes medical and drug claims, as well as eligibility data, for all primary policyholders and beneficiaries. Use of medical services is recorded in the database with date of service, 8 associated diagnoses, performed procedures, billing charges, and payer-reimbursed amounts. Pharmaceutical drug claims are recorded with prescribed medications identified by National Drug Codes, date of prescription filled, days of supply, and payment. This database is compliant with the Health Insurance Portability and Accountability Act. Because the data are commercially available and deidentified, institutional review board approval was not required. Analyses were conducted in September 2016.
Patients were considered eligible if they were continuously enrolled in a health plan during all of 2012. Patients younger than age 18 years (calculated as 2012 − year of birth) were classified as pediatric, and those age 18 years and older were classified as adults. Patient sex was recorded by the insurance company. Each medical claim was associated with up to 2 International Classification of Diseases, Ninth Revision, Clinical Modification (ICD-9-CM) diagnosis codes.
Uveitis (both infectious and noninfectious) diagnoses were identified using the list of ICD-9-CM codes in eTable 1 in the Supplement, modified from the study by Reeves et al.14 All codes in the list from Reeves et al14 were included. An additional 4 ICD-9-CM diagnosis codes were added based on clinical input, which were similar to those included in the Reeves et al14 list and considered relevant for the analysis.
Patients were identified as having uveitis if they had at least 2 medical claims with uveitis diagnoses on separate days in 2012; requiring 2 separate uveitis diagnoses helped to minimize misdiagnoses. Patients with NIU were further identified using the ICD-9-CM codes listed in eTable 2 in the Supplement. Using these same codes, patients were categorized into mutually exclusive groups by location of inflammation and classified as having pan-NIU if they had evidence of NIU in at least 2 different locations (anterior, intermediate, or posterior).
Among patients with anterior NIU, those with severe NIU were identified as those refractory to topical corticosteroid therapy (had either  a continuous treatment episode of topical corticosteroids for at least 90 days, allowing for a gap of 15 days in medication supply, on or after the first anterior NIU diagnosis; or  received step-up therapy for anterior NIU). Step-up therapy was defined as at least 1 prescription of a topical corticosteroid on or after the first diagnosis of anterior NIU, followed by at least 1 prescription of corticosteroids, immunosuppressants, or biologics, with the additional requirement that the patient had no prescription of corticosteroids, immunosuppressants, or biologics before the first prescription topical corticosteroid.
Patients with anterior NIU were classified as having an associated systemic condition if they had any of the following diagnoses during the same year as the diagnosis for anterior NIU (eTable 3 in the Supplement): plaque psoriasis, Crohn disease, ulcerative colitis, psoriatic arthritis, Behçet syndrome, spondyloarthritis, ankylosing spondylitis, undifferentiated spondyloarthropathy, multiple sclerosis, and sarcoidosis (for pediatric patients, the only systemic condition considered was juvenile idiopathic arthritis [JIA]). Patients with anterior NIU without these conditions were classified as idiopathic. Finally, patients with idiopathic anterior NIU were classified as having severe NIU, as described previously.
Prevalence was defined as the number of diagnosed uveitis cases recorded during 2012 per 100 000 adult patients continuously enrolled in a health plan during 2012. Confidence intervals were calculated assuming a Poisson distribution.11 To estimate the number of uveitis cases in the United States using the most recent population estimates, the calculated prevalence was applied to the relevant 2015 US Census population estimate. The 2015 adult US population was estimated to be 247.2 million, calculated as the total 2015 US population (321.4 million15) multiplied by the 2014 adult population proportion (76.9% = 245.3 million/318.9 million16) (at the time of this writing, the most recent data available for US population age breakdown were from 2014). The 2015 US pediatric population size of 74.2 million was estimated using the US 2015 population of 321.4 million multiplied by the 2014 pediatric population proportion of 23.1%.16 The analyses were conducted separately for adults and children and stratified by sex and adult age groups.
The base sample consisted of approximately 4 million eligible patients, approximately 1.9 million men and 2.1 million women (the number of male and female patients in the sample does not sum to the total number of eligible patients because, for a small number of patients in the database, the sex variable was missing; Table 1).
There were 5299 cases of uveitis, a prevalence of 133 per 100 000 (95% CI, 129.1-136.3) (Table 2). Noninfectious uveitis made up 91% of these cases, with a prevalence of 121 cases per 100 000 (95% CI, 117.5-124.3). Women had a higher prevalence of uveitis than men, with 146 cases per 100 000 women (95% CI, 140.3-150.8) and 119 cases per 100 000 men (95% CI, 114.1-123.9). Nonanterior NIU made up 923 of the total NIU cases, with intermediate, posterior, and pan-NIU prevalence of 1.0 (95% CI, 0.8-1.5), 10 (95% CI, 9.4-11.5), and 12 (95% CI, 10.6-12.7) per 100 000, respectively. Women had a higher nonanterior NIU prevalence (27 per 100 000; 95% CI, 24.6-29.1) than men (19 per 100 000; 95% CI, 17.3-21.2). Approximately 81% of NIU cases (3904 cases) were classified as anterior NIU, with a prevalence of 98 per 100 000 adults (95% CI, 94.7-100.9).
Among patients with anterior NIU, 11% were associated with systemic autoimmune conditions, with the remaining 3472 patients (89%) classified as idiopathic. This resulted in an idiopathic anterior NIU prevalence of 87 per 100 000 adults (95% CI, 84.1-89.9). About 10% (n = 392) of all patients with anterior NIU and 9% (n = 326) of patients with idiopathic anterior NIU were considered to be severe cases, for a prevalence of 8 to 10 per 100 000.
Based on the prevalence results presented here, if we were to apply our prevalences to the US population, approximately 328 019 cases of uveitis would be estimated for the United States in 2015 (95% CI, 319 192-336 994), with 298 801 resulting from NIU (95% CI, 290 512-307 324): 130 028 men (95% CI, 124 523-135 712) and 168 574 women (95% CI, 162 346-175 039) (Figure). Among all NIU cases, 57 136 would be cases of nonanterior NIU (95% CI, 53 405-60 822): 23 064 men (95% CI, 20 814-25 506) and 34 010 women (95% CI, 31 225-36 937). Pan-NIU, the most common type of nonanterior NIU, would be expected to include 28 661 cases (95% CI, 26 208-31,400), followed by posterior (25 751; 95% CI, 23 241-28 433) and intermediate (2724; 95% CI, 1978-3709). A total of 241 665 cases would be expected to be anterior NIU (95% CI, 234 140-249 469), and of those, 214 924 idiopathic (95% CI, 207 932-222 272). We would expect 24 266 cases of anterior NIU (95% CI, 22 005-26 702)—or 20 180 cases of idiopathic anterior NIU (95% CI, 18 049-22 499)—to be classified as severe.
Table 3 presents analyses according to 5 age categories for adults: 18 to 30, 31 to 44, 45 to 54, 55 to 64, and 65 years and older. The prevalence of NIU increased with patient age, from 51 cases per 100 000 adults (95% CI, 46.3-56.8) in the 18 to 30 age group to 190 per 100 000 in the 65 years and older age group (95% CI, 180.5-199.3). Across all age groups, the most common location of the uveitis was anterior, with a prevalence of 42, 75, 95, 116, and 155 cases per 100 000 for the age groups 18 to 30, 31 to 44, 45 to 54, 55 to 64, and 65 years and older, respectively. Among the youngest adult age group (18-30 years), pan-NIU was the most common type of nonanterior NIU, comprising 55% of nonanterior cases and having a prevalence of 5 cases per 100 000 (95% CI, 3.8-7.2). For the oldest patients (age 65 years and older), the most common nonanterior location of inflammation was posterior, comprising 53% of nonanterior cases and having a prevalence of 19 cases per 100 000 (95% CI, 15.9-21.9).
Patients in the oldest age group were more likely to have severe anterior NIU than those in the youngest adult group. Only 3% (10 cases) of anterior NIU cases among the youngest patient group were severe, with a prevalence of 1 case per 100 000 patients (95% CI, 0.7-2.5), while the share of severe cases among patients with anterior NIU was 13% (165 cases) in the oldest group, with a prevalence of 20 per 100 000 (95% CI, 16.8-22.9). The share of idiopathic anterior NIU cases remained relatively stable across age groups, ranging from 85% to 93% of anterior NIU cases; the prevalence increased from 39 of 100 000 patients age 18 to 30 years (95% CI, 34.2-43.3) to 143 of 100 000 patients age 65 years and older (95% CI, 135.2-151.5). Based on our modeling, there were an estimated 88 443 patients with NIU older than 65 years in the United States in 2015 (95% CI, 84 140-92 903) (Table 4); 9166 of those represent severe cases of anterior NIU (95% CI, 7831-10 675), while 8278 have both idiopathic and severe anterior NIU (95% CI, 6992-9742).
Pediatric NIU prevalence estimates are reported in Table 2. The sample size of eligible pediatric patients was 932 260 (475 481 boys [51.0%] and 456 767 girls [49.0]; Table 1). Uveitis was reported in 291 cases, with a prevalence of 31 per 100 000 patients (95% CI, 27.7-35.0). The prevalence for boys was 34 per 100 000 (95% CI, 28.8-39.5), while for girls it was 28 per 100 000 (95% CI, 23.8-33.8). Of the total uveitis cases, almost 95% (275 cases) were NIU; boys had an NIU prevalence of 32 per 100 000 (95% CI, 27.5-37.9), while the NIU prevalence for girls was 26 per 100 000 (95% CI, 22.0-31.7). Most male and female pediatric NIU cases had anterior uveitis (207 cases), and a prevalence of 22 per 100 000 (95% CI, 19.3-25.4). Among girls, pan-NIU was the most common type of nonanterior NIU, comprising 74% of nonanterior NIU cases, while among boys, posterior uveitis was the most common nonanterior location of inflammation (49%).
In the pediatric population, the only systemic condition considered was JIA. Overall, 26.2% (72 cases) of pediatric NIU cases were associated with JIA. In anterior NIU pediatric cases, 64 cases were associated with JIA, with a prevalence of 7 cases per 100 000 (95% CI, 5.3-8.8), accounting for 30.9% of the pediatric anterior NIU sample. Meanwhile, 143 cases (69.1%) were not associated with JIA, with a prevalence of 15 per 100 000 (95% CI, 12.9-18.1). Among the male pediatric population, the prevalence of anterior NIU not associated with JIA was 19 per 100 000 (95% CI, 15.4-23.5); among females, it was 11 per 100 000 (95% CI, 8.5-14.9).
Based on our population modeling, we estimated that there were 23 152 pediatric uveitis cases in the United States in 2015 (95% CI, 20 546-25 961), with 21 879 of those being NIU (95% CI, 19 360-24 626) (Figure). We projected 12 272 male pediatric NIU cases (95% CI, 10 419-14 359) and 9613 female pediatric NIU cases (95% CI, 7984-11 504). In 5728 of these NIU cases, the patients also had a diagnosis of JIA (95% CI, 4450-7195). The estimated number of pediatric cases with anterior NIU was 16 469 (95% CI, 14 316-18 840), and of these, 5092 were associated with JIA (95% CI, 3931-6527). An estimated 5410 cases had nonanterior NIU (95% CI, 4228-6824), composed of 318 intermediate (95% CI, 74-816), 2069 posterior (95% CI, 1335-3041), and 3023 pan- (95% CI, 2151-4154) NIU cases.
This study focused on NIU, a subset of uveitis often subsumed by research on uveitis prevalence but not analyzed on its own. However, understanding the prevalence of NIU separately from overall uveitis is crucial, as many patients with NIU, particularly those for whom topical corticosteroids is not effective, lack simple treatment options that do not impose significant health risks. This study found an overall NIU prevalence of 121 per 100 000 adults and 29 per 100 000 children, with a higher prevalence among adult females than adult males and among older individuals.
Although the primary objective of this study was to estimate the prevalence of NIU, we also estimated the prevalence of uveitis overall because it is useful for comparisons with previous research that have found considerable differences in uveitis prevalence depending on the population studied. The combined adult and pediatric uveitis prevalence found in this study (113.5 per 100 000) is similar to the prevalence of 115.3 per 100 000 found in a previous large population-based study, the Northern California Epidemiology of Uveitis Study.12 However, the Northern California Epidemiology of Uveitis Study did not report prevalence separately by uveitis etiology, so comparing NIU prevalence alone is not possible. A 2008 study of Veterans Affairs Medical Center patients in the Pacific Northwest, the Northwest Veterans Affairs (NWVA) Study,13 found a lower uveitis prevalence of 69 per 100 000 (95% CI, 57-83). The lower overall prevalence found in that study could be driven by the low number of female patients in their sample, as both the current study and the Northern California Epidemiology of Uveitis Study found that the prevalence of uveitis was higher in females. The Northwest Veterans Affairs Study reported prevalence by cause of uveitis and found that 80% of cases were noninfectious (infectious causes of uveitis include cytomegalovirus, herpes simplex, herpes zoster, acute retinal necrosis, zoster keratouveitis, histoplasmosis, and unusual postoperative causes). This is slightly lower than in the current study (90%). A third study, the Pacific Ocular Inflammation Study11 found an even lower uveitis prevalence of 57.5 to 58.0 per 100 000, using a population of patients enrolled in the Kaiser Permanente Hawaii Health Plan in 2006-2007. However, as discussed in the Pacific Ocular Inflammation Study, these numbers may be driven by the large proportion of Pacific Islanders in the patient population (approximately 30%) who were found to have significantly lower uveitis prevalence than other demographic groups. The Pacific Ocular Inflammation Study did not report prevalence by cause or for NIU.
There also have been several smaller studies of patients with uveitis that, although not reporting NIU prevalence, shed some light on the relative frequency of NIU compared with infectious uveitis. A study of 927 patients with severe or sight-threatening uveitis in a hospital in France found 68% (n = 630) of cases were noninfectious.17 However, this sample may not be representative, as it focused only on a subgroup with severe uveitis. A 2005 study of 276 children with uveitis in Israel determined a noninfectious etiology in 67% (n = 184) of the cases; of those patients with NIU, 22.3% (n = 41) were found to have JIA as the associated etiology, which is similar to the current study’s finding of 26% (n = 72) of pediatric NIU cases associated with JIA.18 In 2012, researchers looking at a patient population of 2619 children and adults with uveitis in Vienna found 81% (n = 2124) of cases were noninfectious.6
The current study had several advantages and limitations. The large sample provided by the OptumHealth Reporting and Insights database allows for the detailed analysis of NIU separately from infectious uveitis. It provides a large sample of the commercially insured US population, although results may not prove to be generalizable to US patients with no insurance, those with Medicaid, or veterans.
Research using claims data may have possible database errors or omissions. In addition, the identification of uveitis and other conditions in this study was based on ICD-9-CM diagnosis codes and not confirmed by ophthalmologic examination, which could lead to some misidentification and/or misclassification of the uveitis types. However, this risk may be mitigated because we used adjudicated claims, so initial coding errors by the practitioner or insurer may have been corrected, and we required 2 diagnoses on different dates to minimize the risk that these patients’ illnesses were misdiagnosed. Further misclassification may have occurred owing to the study requirement that patients be categorized into mutually exclusive NIU classifications. In particular, patients with evidence of NIU in 2 different locations (anterior, intermediate, or posterior) were classified as having pan-NIU, despite the fact that some of these patients may have anterior and intermediate inflammation and, therefore, would not be classified as having pan-NIU by the Standardization of Uveitis Nomenclature criteria.1 Some diseases associated with uveitis do not have specific ICD-9-CM codes (eg, tubulointerstitial nephritis and uveitis syndrome), and because we were unable to discern the proportion of cases with these associated conditions, we likely overestimated the number of idiopathic uveitis cases. In addition, some patients with systemic disease may be categorized as idiopathic if their systemic disease diagnosis occurred after the end of the 1-year period studied here. However, given the rarity of these conditions, we would expect the estimate of idiopathic uveitis to be within the study’s estimated 95% CIs. Finally, while we classified patients as having severe NIU in part by their use of immunosuppressive or biologic therapies, we cannot be certain whether these drugs were being used to treat NIU or another underlying condition. Therefore, the number of patients with severe uveitis may be overestimated.
In conclusion, this study suggests that NIU affected an estimated 298 801 adults (95% CI, 290 512-307 324) and 21 879 children (95% CI, 19 360-24 626) in the United States in 2015, with an estimated prevalence of 121 per 100 000 and 29 per 100 000, respectively. Most of these cases were anterior uveitis, and about 10% of adult patients with anterior NIU were considered severe cases. Nonanterior NIU had an estimated combined prevalence of 23 per 100 000, with pan-NIU and posterior NIU being the most common. A better understanding of the prevalence of noninfectious uveitis will help to determine the number of patients affected by different types of this disease.
Corresponding Author: Jennifer E. Thorne, MD, PhD, Division of Ocular Immunology, Department of Ophthalmology, Johns Hopkins University School of Medicine, 600 N Wolfe St, Woods 476, Baltimore, MD 21287 (email@example.com).
Accepted for Publication: July 11, 2016.
Published Online: September 8, 2016. doi:10.1001/jamaophthalmol.2016.3229
Author Contributions: Dr Thorne had full access to all the data in the study and takes responsibility for the integrity of the data, study supervision and the accuracy of the data analysis
Concept and design: Thorne, Skup, Tari, Macaulay, Chao, Ganguli.
Acquisition, analysis, or interpretation of data: All authors.
Drafting of the manuscript: Skup, Macaulay.
Critical revision of the manuscript for important intellectual content: All authors.
Statistical analysis: Skup, Macaulay, Ganguli.
Obtaining funding: Chao, Ganguli.
Administrative, technical, or material support: Suhler, Skup, Macaulay, Ganguli.
Study supervision: Suhler, Macaulay, Chao, Ganguli.
Conflict of Interest Disclosures: All authors have completed and submitted the ICMJE Form for Disclosure of Potential Conflicts of Interest. Dr Thorne serves on the advisory boards of AbbVie and XOMA; is a consultant for Gilead, Navigant, and NightstaRx; and has received research support from AbbVie, Allergan, the National Eye Institute, and XOMA. Dr Suhler serves as a consultant to AbbVie, Mallinckrodt Pharmaceuticals, Santen, and XOMA and receives research support from AbbVie, Bristol-Myers Squibb, Clearside, EyeGate, Genentech, pSivida, and the National Eye Institute. Dr Suhler is supported by the Department of Veterans Affairs and has an unrestricted grant from Research to Prevent Blindness. Drs Ganguli, Skup, and Tari are employees of AbbVie Inc and hold AbbVie stock/stock options. Dr Chao is a former employee of AbbVie Inc and holds AbbVie stock/stock options. Dr Macaulay is an employee of Analysis Group Inc, New York, New York, which received consultancy fees from AbbVie for this project. Analysis Group Inc is an economic, financial, and strategy consulting firm that provides services to a variety of companies in the health care industry.
Funding/Support: The research reported in this article was funded by AbbVie Inc.
Role of the Funder/Sponsor: AbbVie Inc had a role in the design and conduct of the study; collection, management, analysis, and interpretation of the data; preparation, review, or approval of the manuscript; and decision to submit the manuscript for publication.
Additional Contributions: We thank Jin Wei, MS (former employee of Analysis Group Inc), for her assistance with implementing some data cleaning and statistical analyses with guidance and instruction by the authors; however, she did not contribute substantially to the study design and did not participate in drafting the manuscript or its content. She also did not review the final results that are presented in the article or contributed to the interpretation of the results. She did not receive compensation.
Create a personal account or sign in to: