Outcome of polymerase chain reaction (PCR) analysis in 100 suspected cases of infectious uveitis

Aqueous humor (AH) samples (150–200 μl) were collected aseptically in a tuberculin syringe with a 30-gauge needle, under aseptic precautions by a single ophthalmologist as an outpatient department (OPD) procedure, and undiluted vitreous humor (VH) samples were obtained by a 23-gauge needle pars plana vitrectomy in an operation theater. The samples were transferred onto pre-sterilized microfuge tubes and stored at − 20 °C for DNA extraction.

Leukocytes of the buffy coat suspended in 100 μl of aqueous were subjected for DNA extraction following the manufacturer’s instructions of QIAGEN DNA extraction kit, Hilden, Germany.

PCR testing was performed for the commonest causative organisms, namely, CMV, HSV type1 and type2, VZV, T. gondii, M. tuberculosis, P. acnes, Eubacterium, and Panfungus genome using previously published primer sequences.

To detect the DNA for bacterial species and fungus, broad-range PCRs were performed targeting16S ribosomal DNA and 18S ribosomal DNA, respectively, in accordance with our previously reported methodology.

The PCR mixture (50 μl) contained 100 mM of dNTP mixture, 10× PCR buffer with 15 mM MgCl2, 1 μM of each forward and reverse primer, and 3 U/μl Taq DNA polymerase. Ten microliters of extracted positive control or test sample DNA was added to the first round PCR reaction mixture. For the second round amplification, 5 μl of the first round product was added to the 50 μl of the PCR mix containing 10 mM of each dNTP, 10× buffer, 1 μM of each forward and reverse primer, and Tag DNA polymerase. Two controls (one reagent control and another one serves as reaction control) were included in each PCR run. The PCR results were considered valid only when the reagent controls were negative and the specific amplified product was obtained with the positive controls. To prevent contamination DNA extraction, PCR cocktail preparation, amplification, and analysis of results were carried out in physically separated rooms. The M. tuberculosis load was estimated in the DNA extracts of test samples using a commercial kit—Geno-sen’s M. tuberculosis Real Time PCR kit (Genome Diagnostics Pvt. Ltd)—and the assay was performed on Rotor Gene (Hilden, Germany) real-time PCR equipment based on Taqman principle.

DNA was extracted from Mycobacterium tuberculosis strain H37Rv, herpes simplex virus (HSV) 1—ATCC VR 733, herpes simplex virus (HSV) 2 ATCC 753167, cytomegalovirus (CMV)—ATCC 169, varicella zoster virus (VZV)—ATCC Oca strain, Toxoplasma gondii— ATCC 50869, Eubacteria (Propionibacterium acnes lab isolate), and fungus (Candida albicans) ATCC 90028. All the standard strains are maintained in the laboratory.

Visualization of PCR product was done by subjecting 10 μl of amplified reaction mixture to electrophoresis on a 2% agarose gel incorporating 5 μg ml of ethidium bromide in 1× Tris-Borate buffer (pH 8.2–8.6) and documented on gel documentation system (Vilber Lourmat, France).

Initial pre-PCR diagnoses were established on the basis of history, clinical findings, and investigation results. The major outcomes considered were the correlation of the pre-PCR diagnosis with the PCR results and change in treatment modality. The sensitivity and specificity of PCR was calculated based on the final diagnoses derived from the clinical course response to treatment and results of ancillary investigations. Positive predictive value (PPV) and negative predictive value (NPV) for PCR testing were also calculated and derived using Bayes’ theorem.

The treatment consisted of anti-tubercular therapy (ATT) for 9 months for ocular tuberculosis, intravenous acyclovir, or oral valacyclovir for 90 days in HSV or VZV retinitis. Oral valgancyclovir was considered for CMV retinitis and oral clindamycin 6 weeks in ocular toxoplasmosis and broad-range antimicrobials in endophthalmitis. Oral corticosteroid was considered whenever required.

Approval from institution ethics committee was taken and adherence to the tenets of Declaration of Helsinki maintained. Descriptive statistics were computed and statistical software (SPSS 14) was used for univariate and multivariate analysis and for the logistic regression analysis; p value < 0.05 was considered to be significant. The 95% bootstrapped confidence intervals were computed for all summary statistics. A chi-square test and paired t test was used to compare sensitivity values. Stepwise regression method was used to determine the significant predictors for final visual acuity. Then, multiple regression models were used to predict final visual acuity with different types of uveitis.

The mean age was 39.2 ± 15.4 years, the range being 9–72 years. The male were 52% and female were 48%. There was unilateral involvement in 82% and bilateral in 18% of the cases. The demographics and the clinical features of the patients are shown in Table 1.

Mean visual acuity at presentation was 0.73 logMar unit and the mean intraocular pressure was 15.56 ± 3.76 mmHg. Among these 100 cases subjected for ocular fluid PCR analysis, posterior uveitis was the most prevalent type (38%) followed by anterior uveitis (34%).The classification of various types of uveitis are given in Table 1. There were 37% patients who had an acute episode of uveitis, 30% with chronic uveitis, and recurrent uveitis in 33%.

Initial pre-PCR diagnoses were established based on the history, clinical findings, and investigation results. The commonest type of uveitis subjected for PCR was posterior uveitis (38%). The commonest suspected infectious etiology was tuberculosis (37.1%) followed by viral uveitis (23.8%) which is shown in Table 2.

Among 100 eyes of 100 patients, PCR analysis confirmed the initial clinical diagnosis in 70.1% patients. The correlation of pre- and post-PCR diagnoses is depicted in Table 2. But PCR altered our treatment in 17.7%. The overall sensitivity and specificity of PCR analysis of ocular fluid was found to be 90.2 and 93.9%, respectively. The PPV, defined as the likelihood of having disease related to the tested infectious agent given positive PCR results, was 93.9% and NPV, defined as the likelihood of not having the specified disease given negative PCR results, was 90.2% in the overall analysis (Table 3). The sensitivity, specificity, and predictive values of the individual organism were also analyzed (Table 4) and the agarose gel electrophotogram of detection of various DNA genomes shown in Figs. 1, 2, 3, and 4. Real-time PCR was done for M. tuberculosis and this showed the quantitative value ranging from 32 to 2722 c/ml.

The visual acuity at presentation was correlated with the visual acuity at the final follow-up. Mean visual acuity at initial visit and final visit was 0.73 and 0.63 logMar unit, respectively, using t test. Though the vision improved, no significant p value change was noted.

The stepwise regression method was used to determine the significant predictors for final visual acuity. Then, multiple regression model was used which revealed intermediate uveitis and panuveitis as important predictors for final visual acuity. The ANOVA test of regression model showed that presence of intermediate uveitis and panuveitis was directly related with decline in final visual acuity by 0.491 + 0.646 logMar unit and 0.491 + 0.983 logMar unit. The coefficient of determination showed R2 = 0.06.9% which meant that only 6.9% of the variation in the dependent variable is explained by the model, while the remaining 93.1% is left unexplained.

The collection of the aqueous fluid was done safely in the outpatient department (OPD) and only one eye had complication in form of hyphema.