Abstract
Purpose :
To evaluate the impact of PCR sampling on diagnosis and treatment of infectious uveitidies at a large tertiary care facility
Methods :
This is a retrospective chart review of patients with at least one PCR sample performed from 2014-2019 at New York Eye and Ear Infirmary of Mount Sinai. At least one follow up visit following results of PCR testing was required for inclusion. If a patient had multiple PCR samples taken, only the first sample was included. The samples were divided into three categories based on pre-PCR diagnosis: those where pre-sampling diagnosis was presumed not infectious (group 1), those where pre-sampling diagnosis was presumed infectious (group 2) and those where pre-sampling diagnosis was concerning for infection but still unknown (group 3).
Results :
108 cases were available for analysis in the study. 25 of the 108 samples yielded a positive PCR result (23%). Majority of patients fell into group 3 where diagnosis was unknown (56%). Presumed infectious and noninfectious cases made up 22% of the population each. Table 1 summarizes treatment and diagnostic changes made in response to PCR sampling results. A chi square analysis was performed. The results of PCR testing had a more significant impact on diagnosis in those cases where pre-sampling diagnosis was unknown versus those where it was presumed infectious or presumed noninfectious (77% vs. 21%; 77% vs. 13%, p<0.0001). In cases where the presumed diagnosis was infectious, the rate of diagnosis change was similar to those where the diagnosis was presumed noninfectious (21% vs. 13%, p>0.05). Similarly, treatment changes were made based on PCR results more frequently in cases where the diagnosis was unknown compared to when an infectious entity is presumed (38% vs. 25%, p<0.0001). There were no cases where treatment was changed in relation to PCR when a noninfectious entity was the pre-sampling diagnosis.
Conclusions :
In our experience, PCR sampling is undertaken in various clinical scenarios, including when an infectious entity is not suspected, when an infectious entity is suspected, and when diagnosis is unknown. As expected, it has the greatest impact on diagnosis and treatment when the clinical diagnosis is unknown. When PCR sampling is confirmatory in nature, it has a minimal impact on diagnosis and treatment, especially when compared to unknown cases.
This is a 2021 ARVO Annual Meeting abstract.