Keywords

These keywords were added by machine and not by the authors. This process is experimental and the keywords may be updated as the learning algorithm improves.

FormalPara Core Messages

  • The epidemiology of uveitis (as well as other diseases) changes over time due to changes in population demographics, environmental factors, infectious disease fluctuations, and treatment evolution.

  • Most of our present understanding of the epidemiology of uveitis is from case series derived from subspecialty referral practices, typically located in academic centers. These data may not reflect the overall spectrum of disease in the general population because of referral bias.

1 Introduction

Uveitis has been said to cause between 2.8 and 10 % of blindness in the USA and cause approximately 30,000 new cases of blindness per year [1, 8, 9, 14, 25, 34, 35]. Even if uveitis does not result in blindness, it can cause visual loss and multiple other ocular complications. In the Northern California Epidemiology of Uveitis Study, 31.9 % of the patient cohort was affected by one or more complications over the 12-month study period (Gritz DC and Wong IG, The Northern California Epidemiology of Uveitis Study, previously unpublished data) thus, uveitis represents an important group of sight-threatening conditions. Uveitis is not only responsible for significant loss of vision and ocular morbidity, but also results in great medical economic costs and personal impact on patients in the form of financial and psychological stress ([33], Gritz DC and Wong IG, unpublished data). Despite the facts, there is a paucity of evidence-based data on the epidemiologic, socioeconomic, and therapeutic aspects of this disease [8]. Thankfully, the number of publications in this area seems to be expanding.

Most of our knowledge about the epidemiology of uveitis is from case series derived from subspecialty referral practices. Typically, these practices are associated with universities or other academic centers. Subspecialty practices and referral centers are readily able to compile case series and help to characterize uncommon diseases, in particular. These case series have helped us to better understand many uveitic diseases, such as Vogt-Koyanagi-Harada disease and Behçet’s disease. Since the most difficult and complicated cases are referred to these practices, observant physicians may describe new syndromes and unusual manifestations of established diseases. However, since only a subpopulation of patients is referred to these practices, a referral bias exists in any case series that is compiled from a subspecialty practice. For example, because the most difficult cases are referred to subspecialists, there is a referral bias which could result in misunderstanding diseases and the true spectrum of disease in the general population.

McCannel and associates examined the difference between general ophthalmology and subspecialty practices [23]. They compared consecutive uveitis cases seen in community-based comprehensive ophthalmology practices and the UCLA uveitis referral practice. There were significant differences in the anatomic location of disease in the two groups of patients with intermediate, posterior, and panuveitis being more common in the referral practice (p = 0.00005). A cause or clinical syndrome could be assigned to a higher percentage of cases in the referral practice compared to the community-based practices (p = 0.03). These statistical differences in two clinical aspects of uveitis are just a glimpse of the differences that exist between community-based and referral practices.

This chapter will discuss epidemiologic methodology, our present knowledge of the epidemiology of uveitis, focusing mainly on population-based studies, and the factors that influence epidemiology. Although the number of population-based studies is limited, these studies yield valuable information about how uveitis affects populations and which populations are at greatest risk. There is a great deal that we do not understand about the epidemiology of uveitis. Although the knowledge we have may point to specific populations at risk, the present studies do not elucidate the reasons why certain populations are at greater risk.

2 Epidemiologic Methodology

Whenever studies are performed, it is important that similar definitions are utilized that have clinical correlation. For instance, in the study of uveitis overall, cases are restricted to intraocular inflammation that is not due to acute trauma, normal postoperative inflammation, postoperative endophthalmitis, and inflammation that is secondary to bacterial or fungal corneal ulcers. Iritis due to herpes simplex or varicella-zoster viruses is typically included in studies of uveitis overall.

There have been efforts to standardize uveitis terminology in the past [4, 20]. The publication by the Standardization of Uveitis Nomenclature (SUN) Working Group was a significant contribution to help clinicians and researchers around the world to “speak the same language” [20]. The SUN Working Group clarified terminology in several areas: anatomic classification, descriptors regarding the onset and course of uveitis, and standardized grading criteria for anterior chamber cells and anterior chamber flare (see Chap. 2).

In regard to specific diseases, Behçet’s disease has had several classification systems from the rheumatologic literature, and there has been an effort to create one standardized system for use internationally [26, 36]. These same classification systems are used by ophthalmologists. The International Uveitis Study Group has contributed to the literature by publishing standardized diagnostic criteria for Vogt-Koyanagi-Harada syndrome [31]. Levinson et al. published the consensus agreement regarding birdshot chorioretinopathy from an international workshop on this disease entity [22]. These publications and the organizational movements behind them are encouraging that epidemiologic research in ophthalmology is on the rise and the methodology behind the studies will become stronger.

Case presentations and case series convey anecdotal clinical experience. The process of tabulating and organizing case series may yield additional understanding about the group of patients observed. Case series are descriptive in character, allowing little appropriate statistical analysis. Comparison of the ratio of men to women in the cases can operate on the assumption that the population is made up for 50 % of each gender; however, this assumption is not always true. In some areas, there are gender inequalities in access or tendency to seek medical care. Because of the differences in life expectancy, there are more females present in older age groups. Statistical analyses of age in case series assume that the population is equally distributed across different age groups. This is not a valid assumption, since age distribution varies markedly between different populations and over time. Care must be taken when drawing conclusions based on anecdotal experience.

Appropriate epidemiologic methods provide a context for comparison. In population-based studies, knowledge of the denominator study population can shed new light on the disease in the study cohort. For example, in the Northern California Epidemiology of Uveitis Study [17], the largest numbers of uveitis patients were in the 25–44- and 45–64-year-old age groups, together comprising 70.1% of the case cohort (Table 4.1). If one based assumptions on anecdotal experience alone, the conclusion might be that uveitis is most common in these two age groups. If one examines the study population, however, these two age groups make up the largest portion of that population, comprising 58.0% of the total population. When the “big picture” is taken into account, the rate of disease is highest in the oldest age group. Although the actual number of cases is small from this group, the number of people in this group is very small, so the rate of disease is actually highest, compared to the other age groups.

Table 4.1 Prevalent cases of uveitis and ratio of population affected by the disease over a 12-month period
Full size table

In hospital- and clinic-based case series, one can change the study design from a simple case series into a case-control study. This allows statistical comparisons of demographics, potential risk factors or associations, and clinical findings. Deciding on an appropriate control group is the challenge of this methodology. If the control group is clinic- or hospital-based, then Berkson’s bias can alter the findings of the study. This states that hospital-based control groups are more likely to have health problems compared to population- or community-based control groups [3]. Thus, Berkson’s bias causes a bias towards the null. Although not a uveitis study, Pratiputibalong et al. observed the demonstration of Berkson’s bias in a case-control study of herpetic eye disease [30]. In this study, atopy was more common in the clinic-based control group than in the population-based control group. This difference affected the statistical comparisons accordingly when these two different control groups were compared to the cases with herpes eye disease, resulting in a bias towards the null when comparing the clinic-based control group and the cases, in contrast to the population-based control group.

3 Present Knowledge of Uveitis Epidemiology

As stated in the introduction to this chapter, our epidemiologic knowledge of uveitis is limited.

The Northern California population studied by Gritz and Wong offered a unique opportunity to examine an ethnically diverse population. The Kaiser Permanente Health Plan is responsible for over one-third of the population of Northern California and includes a cross section of total population. In this study, the disease rates for uveitis were 2.74 times higher than they were in the previous US study [8, 17]. Darrel and colleagues published the first population-based study of uveitis in the literature [8]. This study was performed in Olmsted County, Minnesota, examining a study period from 1945 to 1954. Figure 4.1 compares the age-stratified rates of disease in the two studies.

Fig. 4.1
figure 1

Age-stratified incidence rates of uveitis for the Northern California Epidemiology of Uveitis Study [17] and the Olmsted County study [8]. The rate is shown as the number of cases per 100,000 person-years. The findings in the study were different, with the Northern California study having rates of disease 2.67 times higher than the Olmsted County study (p = 0.001). The increasing rates of disease related to age observed in the Northern California study were statistically significant (p < 0.001)

Full size image

Because of previous studies, the very high rates of disease observed in the Northern California study, especially in people over 65 years old, were initially doubted [18]. However, higher than expected rates of disease have since been observed in the older population [32]. Reeves and associates examined the rates of uveitis in people 65 years old and older in the Medicare National Long-term Care Survey from 1991 through 1999. This study reported an incidence rate of 340.9 cases per 100,000 person-years and a cumulative prevalence ratio of 511 cases per 100,000 persons in 1991, accumulating over the 9-year study period to 1,231 cases per 100,000 persons in 1999. This compares to an incidence rate of 19 cases per 100,000 person-years in those 65 and older in the Olmsted County cohort from 1945 to 1954 [8]. Gritz and Wong found a rate of 102.7 cases per 100,000 person-years [17]. Thus, Reeves and associates found disease rates to be even higher than Gritz and Wong during a similar study period.

When the rates of disease in the Northern California study are age–and sex–stratified, there are higher rates of disease in older women with differences between people with new-onset disease and ongoing (prior-onset) disease (Fig. 4.2). There was a trend towards increased rates of new-onset disease in older women (p = 0.07). The difference in rates for older women is greater in the people whose inflammatory disease began before the study period (p < 0.001). Comparing differences in incidence rates and prevalence ratios can suggest who may be at risk for developing ongoing disease. In this case, it appears that older women may be at a slightly higher risk for a first episode of uveitis and that older women may be at significantly higher risk of developing ongoing disease after that first episode. This is important, since vision loss and ocular complications typically occur with ongoing disease as opposed to a single episode of disease. Additional studies, have not as yet, reported these gender differences.

Fig. 4.2
figure 2

Age- and gender-stratified prevalence rates of uveitis for the Northern California Epidemiology of Uveitis Study [17]. The prevalence ratio of disease observed in women was higher than that seen in men. These differences were influenced by increasing age (p < 0.001). Thus, older women had the highest rates of ongoing uveitis

Full size image

4 Influence of Ethnicity on the Epidemiology of Uveitis

Of the publications that exist on the rates of disease (Table 4.2), most work has been done in populations that are of Northern European origin [8, 24, 27, 33]. The population studied by Darrel and associates was predominantly of Scandinavian descent. More recent studies have been performed in Finland [24, 33] and Switzerland [37]. The results of these studies are shown in Table 4.2. It is interesting to note that these studies, performed in similarly “white” populations, have similar rates of disease, ranging from 17 to 22.6 per 100,000 person-years.

Table 4.2 Results of previous studies on the overall incidence and prevalence of uveitis
Full size table

There is less information on the epidemiology of uveitis in people of color, and only two studies have been published on the rates of disease in populations of color. Freedman examined the rates of uveitis in Bantu-speaking black South Africans [12]. He found an incidence rate of 27.2 per 100,000 person-years. This estimated rate could have been lower than the actual rate of disease because of access to care. This rate, however, is higher than the rates reported in the white populations of Finland and Minnesota. Dadonda and associates performed a survey of eye disease in India and found a point prevalence ratio of 714.3 cases per 100,000 people [7]. This point prevalence ration is dramatically higher than the findings from the studies done in Finland and Minnesota.

In part of the Northern California Epidemiology of Uveitis Study, a representative subpopulation of the study population was evaluated in regard to ethnicity [16]. This revealed a slightly increased incidence of uveitis among people of African descent, compared to those not of African descent (Fig. 4.3; odds ratio = 1.5, p = 0.14). For people with ongoing, prior-onset disease, the difference in rates between people of African descent and others was more striking (Fig. 4.4; odds ratio = 2.1, p < 0.001). In this case, we see a similar relationship between incidence and prevalence in people of African descent and older women in the larger study. So these findings suggest that people of African descent may be at slightly higher risk of having a single episode of disease with a significantly higher risk of developing ongoing disease after that first episode. This fact is not due simply to the group of patients who have uveitis due to sarcoidosis. People of African descent were at a much greater risk of having either biopsy-proven systemic sarcoidosis or “ocular” sarcoidosis (clinical findings suggestive of sarcoidosis, accompanied by laboratory or radiographic findings consistent with sarcoidosis), as shown in Table 4.3. However, in people with uveitis where no etiology or syndrome could be found to be underlying the uveitis, the rates of disease continued to be highest in people of African descent (Table 4.4). Table 4.4 shows the distribution of people with a complete work-up and incomplete work-up for whom no diagnosis was found for the uveitis. There was no difference in the rates of complete and incomplete work-ups between different ethnic groups. There were higher rates of uveitis in both of these diagnostic categories for people of African descent. A complete work-up was based on the clinical judgment by two uveitis specialists, after reviewing the chart in a masked fashion. The rates of uveitis for Caucasians remained significantly higher than previous studies in the USA and Northern Europe.

Fig. 4.3
figure 3

Incidence rates for uveitis among different ethnic groups for the Northern California Epidemiology of Uveitis Study [17]. In this study of a subpopulation of the entire Northern California study, involving 2 communities, the overall rate of uveitis was 50.0 cases per 100,000 person-years, which was not statistically significantly different from the rates of disease seen in the larger study (52.4 cases per 100,000 person-years; p = 0.64). People of African descent had a trend towards higher rates of uveitis compared to people of non-African descent (OR = 1.5; p = 0.14)

Full size image
Fig. 4.4
figure 4

Rates for active, ongoing uveitis for those with onset of disease prior to the among different ethnic groups for the Northern California Epidemiology of Uveitis Study [17]. In this study of a subpopulation of the entire Northern California study, involving 2 communities, the overall ratio of people with ongoing uveitis was 116.4 cases per 100,000 person, which was not statistically significantly different from the rates of disease seen in the larger study (115.3 cases per 100,000 person-years; p = 0.9). People of African descent had higher rates of uveitis compared to people of non-African descent (OR = 2.1; p < 0.001)

Full size image
Table 4.3 Differences in frequency and rates of sarcoidosis-related uveitis for people of African descent versus non-African descent
Full size table
Table 4.4 Differences in frequency and rates of idiopathic uveitis for people of African descent versus non-African descent
Full size table

The reasons for this apparent increased risk for people of African descent are unclear at present. In a study examining the relationship of ethnicity and inflammatory bowel disease, the rates of Crohn’s disease were similar among African Americans and whites [2]. However, African Americans with Crohn’s disease had a significantly higher incidence of associated arthritis (p = 0.004) and ophthalmological manifestations, notably uveitis (p = 0.028). These data are interesting in light of the findings in the Northern California study, suggesting a possibility that there may be an increased risk for ocular manifestations of autoimmune disease affecting people of African descent. Further research will hopefully improve our understanding.

5 Factors that Influence Changing Epidemiology Over Time

There are multiple factors which can influence the epidemiology of uveitis over time. Just as geographic location can influence epidemiology, modern migration also affects the types of diseases affecting people. This migration is now occurring within regions, as populations continue to shift from rural to urban areas, as well as between countries. The increasing diversity resulting from migration can shift the epidemiology of disease. This is likely one of the factors influencing the Northern California study compared to the Olmsted County study ([8, 17], Gritz DC and Wong IG, The Northern California Epidemiology of Uveitis Study, previously unpublished data.).

The world’s population is aging and people are living longer. With this aging, the high rates of uveitis in those over 65 years old could have a significant impact on people’s lives, considering the burden of visual loss, ocular morbidity, and economic impact.

Uveitis has multiple etiologies, including autoimmune and infectious etiologies. Factors that influence autoimmune disease and infectious diseases will thus impact the evolution of uveitis epidemiology. There is evidence that other autoimmune disease rates may be increasing [11, 15, 21, 28, 29]. There is also evidence that there may be a relationship between rates of autoimmune disease and rates of infectious disease [6, 38, 39, 40]. There has not been evidence of these relationships yet in ophthalmic diseases.

For uveitis associated with infectious diseases, there is a great potential for impact from epidemics and rising rates of infection, as well as from new and more effective treatments. A dramatic example of this occurred in the past 25 years. The emergence of HIV resulting in the AIDS epidemic of the 1980s affected the types and manifestations of diseases seen in the ophthalmology clinic. Cytomegalovirus retinitis and Kaposi’s sarcoma changed from rare entities to common reasons for visits in ophthalmology clinics, especially in areas with high rates of AIDS. The subsequent emergence of HAART therapy for HIV disease dramatically affected the types and frequency of inflammatory eye disease seen in the ophthalmology clinics [13]. Cytomegalovirus retinitis is again uncommonly seen, except in people who have immunosuppression either from untreated HIV disease or other causes. With the advent of effective HIV therapies, the entity of immune recovery uveitis came to be known [19] (see Chap. 111).

Other therapeutic advances may be affecting the epidemiology of uveitis, such as the effect of antitumor necrosis factor (anti-TNF) agents on anterior uveitis associated with ankylosing spondylitis [5]. It will be interesting to see whether the increasing use of anti-TNF agents in juvenile idiopathic arthritis associated anterior uveitis will affect this uveitic disease’s epidemiology.

Methotrexate alone has been shown to decrease the incidence of anterior uveitis in children with juvenile idiopathic arthritis (JIA) [28]. In a retrospective cohort of JIA patients, those treated with methotrexate had a lower incidence of anterior uveitis compared to patients who were not treated with methotrexate (10.5% vs. 20.2%, respectively, P = .049). Since the cohort was reviewed retrospectively, there were likely differences in the two groups based on the severity of JIA activity. If the incidence was so much less in the methotrexate-treated group, which had more severe disease, the actual effect of methotrexate is likely to be even greater.

Another human T-cell lymphocyte virus, type 1, is now considered to be a global epidemic and has uveitic manifestations [10]. This virus has been linked to adult T-cell leukemia/lymphoma (ATLL) and HTLV-I-associated myelopathy/tropical spastic paraparesis (HAM/TSP), as well as uveitis. Coinfection with both HTLV-I and HIV has been shown to accelerate the progression of AIDS [10]. The emergence of HTLV-I diseases will likely affect the epidemiology of uveitis over time (see Chap. 112).

A study by Wakabayashi and associates found that in Japanese patients there appeared to be a decreased frequency of Behçet’s disease and an increased frequency of tuberculosis [40]. As tuberculosis overall and drug-resistant tuberculosis spread in populations around the word, this could likely affect the numbers of tuberculosis-related uveitis that are seen.

Thus, the epidemiology of uveitis continues to evolve and change. Thankfully, our research methodology seems to be changing and improving, too. In the coming years it will be interesting to see how our understanding of epidemiology and the underlying mechanisms involved improves.

Take-Home Pearls

  • The incidence and prevalence of uveitis-related intraocular inflammation in the USA appears to be higher now than in the past.

  • The incidence and prevalence of uveitis appears to increase with age, particularly in women.

  • There appear to be higher rates of uveitis in African Americans.

  • By thinking more broadly when designing research and putting together case series, better understanding can result. One example is to improve on a simple case series by adding a control group, resulting in a case-control study.