Clinical utility of aqueous humor polymerase chain reaction and serologic testing for suspected infectious uveitis: a single-center retrospective study in South Korea

Uveitis, an important cause of visual impairment in developed countries, affects approximately 200 per 100,000 individuals and accounts for up to 10–35% of severe visual impairment cases [1, 2]. Infectious uveitis comprises approximately 10–20% of all uveitis cases [3, 4]. The common pathogens implicated in infectious uveitis are cytomegalovirus (CMV), herpes simplex virus (HSV) types 1 and 2, varicella-zoster virus (VZV), and Toxoplasma gondii [5‐7]. Thus, early detection of the causative pathogen and appropriate antimicrobial therapy can prevent visual impairment [1].

Several diagnostic tools such as serologic tests, electron or light microscopy, immunoblots, cell cultures, enzyme-linked immunosorbent assay, and Goldmann-Witmer coefficient are available; however, the initial diagnosis of infectious uveitis is mainly based on clinical features alone. Occasionally, such diagnoses can be quite challenging because not all patients present with pathognomonic clinical features of uveitis. Moreover, a miotic pupil or media opacity can mask the pathognomonic features, or an overlap in the phenotypic expression caused by different pathogens can limit the diagnostic capability based on clinical examination alone [8]. Incorrect diagnoses could delay the administration of targeted treatment, thereby resulting in visual impairment, use of drugs with undesirable side effects, and the occurrence of uveitis-related complications [9].

Polymerase chain reaction (PCR) is a fast and reliable method, which can identify common pathogens causing uveitis [10]. Aqueous humor PCR can precisely detect small quantities of pathogenic DNA or RNA [11, 12]. The usefulness of PCR in diagnosing infectious uveitis has been established; however, only a few studies have compared the results of PCR with those of plasma serologic tests. This study aimed to assess and compare the clinical value of aqueous humor PCR and serologic tests in patients diagnosed with suspected infectious uveitis.

Of the 358 patients included in this study, 34 had anterior uveitis, 10 had intermediate uveitis, 22 had posterior uveitis, and 292 had panuveitis. The mean age of the participants was 54.05 ± 16.58 (range, 13–84) years. The overall diagnostic positivity of PCR and IgM serologic tests for infectious uveitis was 28% (99/358 patients) and 11% (38/358 patients), respectively (Table 1). The pathogens identified by PCR were 3 cases of HSV, 17 of VZV, 67 of CMV, 2 of EBV, and 10 of T. gondii. IgM serologic tests identified 13 cases of HSV, 12 of VZV, 7 of CMV, 4 of EBV, and 2 of T. gondii (Table 1).

Ophthalmic findings of HSV-1/HSV-2- and VZV infection included elevated intraocular pressure associated with acute iritis, stellate keratic precipitates throughout the corneal endothelium, large granulomatous keratic precipitates, sectoral or non-sectoral iris transillumination defects, iris pigment epithelium atrophy, decreased corneal sensation, and mydriatic or corectopic pupil at rest. Ophthalmic findings of CMV infection included unilateral (occasionally bilateral) anterior chamber inflammation associated with iris sectoral defects, episodes of ocular hypertension, and diffuse linear or coin-shaped keratic precipitates, occasionally with focal endotheliitis. Ophthalmic findings of Toxoplasma infection were partial- or full-thickness necrotizing retinitis adjacent to an old hyperpigmented chorioretinal scar associated with focal arteritis, overlying vitritis, and anterior chamber cells [15, 16].

The sensitivity, specificity, and the negative and positive predictive values for PCR and IgM test were evaluated. For PCR test, the sensitivity was 0.431, specificity was 0.985, and the negative and positive predictive values were 0.506 and 0.980, respectively. For IgM test, the sensitivity was 0.151, specificity was 0.970, and the negative and positive predictive values were 0.403 and 0.895, respectively.

For cases with clinical diagnosis, the PCR positivity values were 0.158, 0.630, 0.788, and 0.435 for HSV, VZV, CMV, and T. gondii, respectively, whereas the IgM positivity for clinically diagnosed cases were 0.684, 0.444, 0.082, and 0.087 for HSV, VZV, CMV and T. gondii, respectively (Table 2).

This study reports an overall diagnostic positivity of 28% for PCR compared with 11% for IgM serologic tests to identify uveitis-associated infectious agents. A gold-standard diagnostic tool to detect infectious uveitis is lacking despite numerous studies. Clinical diagnosis based on medical history and ocular examination remains the widely accepted methodology; however, few supportive diagnostic tools exist that help to confirm the primary clinical diagnosis.

Aqueous humor PCR has been proposed as a possible diagnostic tool in several studies [7, 10, 12, 17, 18]. However, the usefulness of PCR in the diagnosis of infectious uveitis remains controversial. A retrospective study reported a 10% diagnostic positivity of aqueous humor PCR for diagnosing anterior uveitis [17], whereas another study has reported a diagnostic positivity of 30% [12]. These diverging results can be attributed to different study designs and the type of uveitis [17, 19]. The geographical region and patient population can also affect study results based on epidemiologic variabilities in the spread of viruses. Furthermore, some institutions perform routine PCR for all patients with uveitis, while others decide on a case-to-case scenario based on clinical suspicion [17, 19].

In South Korea, there are some reports on the clinical patterns of uveitis but none on the prevalence of pathogens identified in infectious uveitis cases. In this study, the pathogens detected using either PCR or serologic tests were CMV in 69 (59%) patients, VZV in 23 (20%), HSV in 14 (12%), and T. gondii in 10 (9%) (Fig. 1). CMV and VZV were the causative pathogens in the majority (79%) of infectious uveitis cases. These results differ largely with those reported in other studies on non-Asian Caucasian patients in which HSV, VZV, and T. gondii were the most commonly identified pathogens [20‐22]. Some geographically proximal countries to South Korea, such as India, Africa, and Japan, have reported HSV, T. gondii, and Mycobacterium tuberculosis as the most commonly detected pathogens in patients with uveitis [22‐26].

In this study, EBV was detected by both PCR and IgM serologic tests in two patients with panuveitis and only by IgM serologic tests in two patients with acute retinal necrosis (ARN), one of whom had a co-infection with CMV, and CMV caused the primary infection. In two patients with clinically diagnosed VZV-uveitis, laboratory test results revealed dual infection (EBV + VZV). EBV is considered to infect the ocular pigment epithelial cells [27], but some studies consider EBV as a secondary factor in ocular inflammation rather than as the primary infectious cause [12, 28]. A few studies have reported that EBV infection might result in ARN; however, this association remains controversial [29]. Therefore, more evidence is required to clarify the role of EBV in uveitis.

For IgM testing, there are two aspects that need to be considered. First, despite low sensitivity, the specificity of IgM test is relatively high, with a low false positive value (sensitivity was 0.151, specificity was 0.970, and the negative and positive predictive values were 0.403 and 0.895, respectively, in our study). Second, as there are endemic areas of viral infection, patients might be broadly positive for IgG; this generalized IgG positivity might not provide any evidence for diagnosis for acute infectious uveitis. For example, more than 90% of Koreans are positive for anti-CMV IgG [30]. Overall, we think it is clinically significant to compare the diagnostic value of IgM and PCR, as both diagnostic tests have their own distinct role as an adjuvant diagnostic tool in infectious uveitis.

In the real world, clinical features are always important in establishing prompt diagnosis for appropriate management of infectious uveitis, as early detection of the causative pathogen and appropriate antimicrobial therapy are critical in preventing visual impairment from infectious uveitis. Concurrently, detection of viral DNA from either aqueous humor or vitreous is necessary for final confirmation of clinical diagnosis. In cases where the PCR results were negative, but infectious etiology of uveitis is highly suspected, we tried to obtain more diagnostic clue from the serologic testing.

This study had some limitations. First, this was a single-center retrospective study and was limited to a specific patient population that visited a tertiary, referral-based university hospital located in the capital of South Korea. These factors could have introduced potential bias in the study group. To indicate a more representative Korean population for studying infectious epidemiology in uveitis, we recommend additional multicenter large-sample studies. Second, PCR was performed only when an infectious etiology was suspected based on clinical findings of uveitis. This could have underestimated the diagnostic value of PCR. Third, PCR was not performed for all etiologic agents in a patient because of the small volume of aqueous humor samples, laboratory limitations, and financial burden on the patient. Only the most probable viral markers were tested, and this might have further underestimated the diagnostic value of PCR. Finally, a number of patients had unclear diagnosis because of negative PCR and serologic test results, and in whom ophthalmic findings were used as the gold-standard method for clinical diagnosis (Table 3).