Aqueous humor polymerase chain reaction in uveitis – utility and safety

Background To study the value and safety of aqueous humor polymerase chain reaction (PCR) analysis for Herpes simplex, varicella zoster, cytomegalovirus, Epstein-Barr virus and Toxoplasma gondii in patients with uveitis. Methods Records of 45 consecutive patients with anterior and posterior uveitis who underwent AC paracentesis with PCR were reviewed. The main outcome measure was frequency of PCR positivity. Secondary outcomes were alteration of treatment, safety of paracentesis, and correlation of keratitic precipitates with PCR positivity, Results The overall PCR positivity was 48.9 % (22/45). Therapy was changed because of the PCR results in 14/45 patients (37.7 %). One patient experienced a paracentesis related complication (1/45, 2.2 %) without long-term sequelae. Conclusion Aqueous PCR altered the diagnosis and treatment in over a third of our patients and was relatively safe. Aqueous PCR should be considered for uveitis of atypical clinical appearance, recurrent severe uveitis of uncertain etiology, and therapy refractory cases.


Background
Rapid analysis of aqueous humor by polymerase chain reaction (PCR) in uveitis can be very helpful in leading to a specific diagnosis of infectious etiology and efficient treatment [1][2][3][4][5][6][7]. However, aqueous humor PCR analysis is not routine and remains controversial because of a low proportion of therapeutic change based on the PCR results in some studies, invasiveness of paracentesis bearing potential complications, and false negatives and positives [5][6][7]. On these grounds recent reports even question its utility and discourage its practice in anterior uveitis [5,8]. Recent reports discourage aqueous PCR for the diagnosis of infectious uveitis or favour PCR of vitreous over aqueous [5,9,10].
Although in most cases the working diagnosis of uveitis is based on history and clinical signs, supported by routine blood tests and radiological examinations, diagnostic error can cause increased ocular morbidity and establishing a definitive etiology often can be challenging. Quick identification of infectious etiology is particularly important in management of posterior uveitis because fundus lesions caused by viruses, bacteria, fungi or parasites cannot always be clearly distinguished, especially in the presence of opaque vitreous. Also, the significant proportion of uveitis characterized as idiopathic might be reduced if PCR testing were performed [11].
In this study we report the results of aqueous PCR analysis for infectious agents and the rate of treatment change based on the PCR results in patients with anterior, posterior, intermediate uveitis, and panuveitis, as well as the complication rate of paracentesis. We also analyze the association of keratic precipitates with PCR positivity and suggest guidelines for performing a safe anterior chamber paracentesis at the slitlamp.

Methods
This study was carried out following institutional guidelines and ethical standards of the 1964 Declaration of Helsinki and its later amendments and was approved by the local ethical committee of the Geneva University Hospitals and School of Medicine (Commission cantonale d'éthique de la recherche CCER) (No. . A retrospective review of consecutive patients with anterior, posterior, intermediate uveitis and panuveitis who underwent aqueous PCR analysis was performed by cross-referencing the PCR test list of the Institute for Immunology and Microbiology with the clinical database of the Department of Ophthalmology from January 1, 2011 to March 1, 2015 and yielded 45 uveitis patients (45 eyes).
The 45 patients' records were reviewed for initial working diagnosis, disease course, signs and degree of ocular inflammation prior to paracentesis, clinical rationale for paracentesis, intraocular pressure, PCR results, and paracentesis complications. The grading and classification of uveitis followed the Standardization of Uveitis Nomenclature (SUN) recommendations [12]. In all cases, diagnostic AC paracentesis was performed at the slit lamp following topical anesthesia, disinfection with povidone-iodine, and placement of a sterile lid speculum. A 27-or 30-gauge needle on a 3 ml syringe was used to extract about 0.2 ml of aqueous humor. Semi-quantitative real-time PCR (rt-PCR) was performed within 24 h of sample collection for Herpes simplex (HSV1/2), varicella zoster virus (VZV), cytomegalovirus (CMV), Epstein-Barr virus (EBV) and Toxoplasma gondii DNA. Nucleic acid from aqueous humor was extracted using the EasyMag Nuclisens® (bio-Mérieux Inc., Durham, NC, USA) (100 μl eluted in 25 μl respectively) and used for rt-PCR with an in-house validated method on a StepOne Plus ABI® machine (Thermo Fisher Scientific Inc, Waltham, MA USA). A signal detected during the first 40 cycles revealed the presence of foreign DNA. The primers and probes sequences used for the detection of HSV 1/2, VZV, CMV, EBV and Toxoplasma gondii DNA are given in Table 1.
The primary outcome measure was the frequency of PCR positivity. Secondary outcomes were change in treatment based on the PCR results, safety profile of diagnostic AC paracentesis, and correlation between keratic precipitates (KP) and positive PCR results.
Descriptive statistics were applied to continuous data while categorical data were assessed using IBM SPSS statistical software version 20 (SPSS, Inc, Chicago, Illinois, USA).

Results
Demographics of the 45 patients are presented in Table 2. Five of the 45 patients underwent repeated paracentesis (average 2.6) because of a negative PCR result in patients strongly suspected of having infectious uveitis, so there were 53 samples in total. Mean age was 43.6 ± 18 years with a female to male ratio of 1:1.6. Indications for paracentesis were: 6.7 % (3/45) recurrent anterior uveitis,  Table 3 outlines the overall results and subgroup analysis for PCR positivity. In 21 of the 45 patients (six anterior uveitis, 11 posterior uveitis, one intermediate uveitis, three panuveitis) sampling occurred without previous anti-inflammatory, antiviral treatment. Seven patients (three anterior uveitis and four posterior uveitis) were already under anti-inflammatory therapy at the time of sampling which did not change. Five patients (three anterior uveitis and two posterior uveitis) were already on anti-infectious before sampling, which changed following the result of PCR analysis. Three posterior uveitis, four panuveitis and five anterior uveitis had no antiinfectious therapy prior to PCR testing but then received it after PCR analysis.

Overall PCR positivity
In total, 22/53 samples were positive for presence of foreign DNA (41.5 %). Two samples from one patient drawn within 48 h were positive for the same pathogen and counted only once in the analysis. 7/45 patients (15.6 %) were immunosuppressed at the time of paracentesis, 5 with Adamantiadis-Behçet vasculitis and already on corticosteroids, 1 suspected to have Birdshot chorioretinopathy, and 1 with bilateral posterior uveitis and occlusive vasculitis of unknown etiology.
The presence of foreign DNA was assessed by semi-quantitative rt-PCR. Table  1 demonstrates the PCR primers and probes sequences for HSV, VZV, CMV, EBV and Toxoplasma gondii that were used

Safety profile
Only one patient developed what appeared to be a paracentesis-related complication, increased anterior chamber inflammation (2.2 %) and was hospitalized for 24 h without long-term sequelae.

Discussion
Diagnostic uncertainty and dilemmas are common in the management of uveitis. In this study, aqueous PCR analysis was of considerable diagnostic value with 48.9 %   of patients positive for infectious DNA leading to a change in diagnosis and treatment in over a third of the patients (37.7 %), all of whom responded favorably to the new treatment. It is likely that with a larger test battery for foreign DNA, the yield of detection of specific infectious pathogens would be even greater. PCR from ocular fluids has a very high reliability with a very low false-positive rate [13][14][15][16]. False negative results are more difficult to verify as there is no gold standard for the accuracy of PCR and viral or Toxoplasma gondii cultures are rarely performed [17]. Aqueous PCR analysis in uveitis has been previously reported [5-7, 15, 18, 19] [5,7].
At present, with limited PCR testing, infectious causes represent around 20-30 % of all uveitis with Herpes viruses the most common cause of anterior uveitis in Western countries [20][21][22][23]. Our subgroup analysis supports this finding. A viral etiology was particularly frequent in hypertensive anterior uveitis with 85.7 % (6/7) of patients positive for CMV, VZV or HSV. The presence of an infectious agent was also confirmed in two cases of suspected FUS found to be positive for the Herpes virus genome and who were treated accordingly. In our posterior uveitis patients, we observed a high prevalence of Toxoplasma gondii (10/45, 22.2 %) and lower rates of CMV, HSV and VZV positivity compared to Asian and US studies which reported CMV or CMV and HSV as leading infectious causes followed by Toxoplasma gondii [6,19]. Previous systematic surveys on uveitis from Western Europe confirm Toxoplasmosis as the most common cause of infectious posterior uveitis which could be related to local customs of consumption of very rare meat and raw chopped steak [23][24][25]. Aqueous PCR testing for Toxoplasma gondii seems to be reliable and not influenced by the interval between symptom onset and paracentesis or host immune status but rather by the total size of the retinal lesions [18,26]. Furthermore aqueous PCR testing can be particularly useful for excluding an infectious aetiology in immunocompromised or older patients as the clinical presentation can be atypical and misleading [15,[26][27][28][29], in severe anterior uveitis cases with no fundal view, in severe presumed HLA-B27 uveitis [30] and in cases of presumed infectious posterior uveitis with atypical and sight-threatening inflammation where a rapid and sensitive diagnosis from prompt sampling for PCR analysis is crucial.
Three patients, one PSS, one posterior uveitis, and one panuveitis, were double positive, for VZV and EBV, EBV and Toxoplasma gondii, and HSV and Toxoplasma gondii, respectively, with weak signals for both EBV and Toxoplasma gondii. In these double positive cases, therapy was influenced by clinical criteria with close followup. Double PCR positivity, which can lead to diagnostic challenges, has been previously reported in immunosuppressed or otherwise immunocompromised patients and may be caused by dormant latent viruses or parasites activated secondary to acute infectious uveitis [6,[31][32][33].
In the particular case of EBV, a major limitation is possible cross reactivity with B lymphocytes containing EBV genome in a latent phase, which can lead to false positive results [13]. This phenomenon of dual positivity has been observed in previous studies and in particular in immunocompromised patients; in those cases PCR testing of ocular fluids helps substantially, testing for multiple pathogens before taking the final decision for treatment based on clinical appearance as the phenotype of the intraocular inflammation can be misleading in such cases.
The complication rate of anterior chamber paracentesis in uveitis is low as corroborated in our study [34]. Paracentesis performed at the slit lamp has a low complication rate if: 1) collaborative adult patients with a deep anterior chamber are selected; 2) the eye is properly prepared; 3) a speculum is used; and 4) careful follow-up is possible. Otherwise, paracentesis supine using an operating microscope and appropriate sedation or anesthesia may be indicated. In our experience, it is essential that the hand holding the syringe be braced securely on the patient's face and not on the frame of the slit lamp and that the patient's head rests firmly on the head and chin rests to reduce head movement. Informed consent is required including as potential complications endophthalmitis, corneal abscess, hyphema, conversion to vitreous sampling, and trauma to the cornea, iris, lens, and posterior segment.
Our study has several limitations. First, it was performed in a tertiary care center including uveitis patients in whom there was a strong suspicion for infectious etiology creating a possible referral and selection bias. Furthermore, different geographic locations have different frequency/prevalence of uveitis pathogens and PCR detection thresholds may vary accordingly between studies [5,19,35]. While our study demonstrates a 48.9 % positivity for foreign DNA, the use of antivirals prior to paracentesis may have reduced the virus load below detectable levels in some patients. Finally, although FUS has been strongly associated with rubella virus, it was not tested in our study [36]. One further pathogen that was not tested with PCR but which merits consideration in relevant clinical settings is Treponema pallidum [37]. Although Treponema pallidum infections were considered almost eradicated in developed countries, epidemiologic studies confirm steady increase of primary syphilis infections with a considerable percentage of secondary, ocular manifestations requiring prompt diagnosis and treatment [38,39].

Conclusion
Aqueous PCR analysis is easy to perform, safe, and valuable, establishing a correct etiologic diagnosis and change in treatment in about a third of selected uveitis cases. We consider aqueous PCR most useful for uveitis with atypical clinical appearance, recurrent severe uveitis of uncertain etiology, and treatment-resistant cases. In suspected infectious uveitis where retinal examination may be limited by vitreous opacification prompt intraocular fluid sampling is necessary and important in reaching a preliminary diagnosis and administering therapy. Regardless, empiric treatment should be initiated based on the clinical picture in severe, sight-threatening uveitis cases with no delay; this can be further adjusted depending on PCR results; negative results can be useful in excluding infectious etiology. In the future, a larger battery of PCR tests for infectious agents will probably become routine for diagnosing infectious uveitis with even better accuracy.