Abstract
Purpose :
The diagnostic performance of polymerase chain reaction (PCR) in a general population of patients with possibleToxoplasma gondii chorioretinitis is uncertain. This retrospective, observational study aimed to determine the effectiveness of analyzing intraocular fluid with T. gondii PCR in patients with undifferentiated chorioretinitis.
Methods :
The results of 187 samples tested with PCR for T. gondii and the associated clinical diagnoses were reviewed. Sensitivity, specificity, positive predictive value, and negative predictive value were calculated. Clinical data for PCR positive samples were reviewed.
Results :
PCR was used to analyze 187 samples from patients with possible T. gondii chorioretinitis. Twenty-five samples from 25 unique patients were positive for T. gondii (25/187, 13.4%). There was no relationship between type of fluid analyzed (aqueous or vitreous) and a positive result (p = 0.73), and 52% of PCR positive patients were immune competent. The final clinical diagnosis was T. gondii chorioretinitis in 36 cases, yielding a prevalence of 19.3%. The test characteristics for T. gondii PCR were: specificity 100%, sensitivity 69.4%, positive predictive value 100%, negative predictive value 93.2%.
Treatment changed in 60% of PCR positive cases. Oral Trimethoprim-sulfamethoxazole was the most common treatment (n= 15, 60%), followed by Clindamycin (12, 48%) and Azithromycin (6, 24%). Testing for other patogens were conducted at the discretion of the physician, and two patients in the T. gondii PCR positive cohort had positive results for a second pathogen. One patient was PCR positive for both Herpes Simplex Virus 2 and T. gondii. A second patient was PCR positive for T. gondii and also positive for Treponema Pallidum antibodies.
Conclusions :
PCR testing for T. gondii is a useful diagnostic tool regardless of patient immune status or ocular fluid analyzed. It can often make the diagnosis of toxoplasmosis in the absence of typical fundus findings, with high positive and negative predictive values.
This abstract was presented at the 2019 ARVO Annual Meeting, held in Vancouver, Canada, April 28 - May 2, 2019.