Accuracy of Gram-stained Smears as Screening Tests for Neisseria gonorrhoeae: A Systematic Review and Meta-analysis

Background and Objective: A total of 86.9 million persons worldwide are infected with Neisseria gonorrhoeae (Ng). Although Gram-stained smears (GSS) provide a timeand cost-saving alternative to conventional laboratory tests, their global uptake partly depends on their performance. This study aimed to meta-analyze the diagnostic accuracy of GSS to screen for Ng. Materials and Methods: A literature search was conducted using the MEDLINE (1980 to 2020). Studies were included if they employed GSS to detect Ng in humans and compared the results with reference tests. Results: Eleven studies were reviewed and meta-analyzed and stratified by specimen type (Gram-stained urethral smears and Gram-stained endocervical, urethral swabs and urine smears.) and reference test type (culture method or NAAT). Sensitivity was similarly high in GSS versus NAAT (93% [CI, 64% to 99 %]) and GSS versus culture methods (87% [CI, 74% to 94%]), followed by Gram-stained urethral smears (97% [95% CI, 86% to 100%]) and Gram-stained endocervical, urethral swabs and urine smears (81% [CI, 67% to 90%]). Specificity was also high in GSS versus culture methods (98% [CI, 95% to 100%]) and GSS versus NAAT (94% [CI, 73% to 99%]), followed Gram-stained endocervical, urethral swabs and urine smears (98% [CI, 93% to 99%]) and Gram-stained urethral smears (96% [CI, 78% to 99%]). Conclusions: Data suggest that GSS have the highest accuracy when investigated against reference culture methods, and Gram-stained urethral smears have the highest accuracy, followed by Gram-stained endocervical, urethral swabs and urine smears. Given their accuracy, convenience, and quick turnaround time, GSS may be useful in expanding first-line screening Ng.