COVID-19 testing uses different types of specimens, including nasopharyngeal swabs, saliva, and serum. While the nasopharyngeal swab (NPS) remains the gold standard for diagnosing COVID-19, its invasive nature can cause patient discomfort and requires trained personnel for sample collection. This study aimed to evaluate the effectiveness of nasopharyngeal, serum specimens, and saliva in detecting COVID-19 and to compare saliva with the other two methods. A systematic search was conducted following the PRISMA 2020 guidelines across PubMed, the Cochrane COVID-19 study register, and the Saudi Digital Library. The QUADAS-2 tool was used to assess study quality. The primary outcome measured the sensitivity and specificity of serum, saliva, and NPS, while the secondary outcome focused on comparing the diagnostic accuracy of saliva versus NPS and serum. Data were collected from 39 studies in 20 countries, analyzing 20,024 patients and 22,123 samples. Despite significant heterogeneity (P < 0.001), the meta-analysis revealed significant differences in sensitivity among all specimen types, especially between NPS and saliva. The area under the curve (AUC) values indicated a high diagnostic performance: serum (AUC = 1.00) showed the highest efficacy, followed by saliva (AUC = 0.97) and NPS (AUC = 0.94). These findings suggest that saliva presents a viable, non-invasive alternative for the diagnosis of COVID-19 with comparable reliability to NPS.
