High ferritin serum levels can be found in patients with macrophage activation syndrome, and increased serum ferritin due to cytokine storm have been reported in severe COVID-19 patients. Saliva is being increasingly used in COVID-19 tests as a diagnostic sample for virus detection and quantification. This study aimed to evaluate the possible changes in ferritin in saliva in COVID-19 patients. In addition, the effects of different inactivation SARS-CoV-2 treatments in ferritin measurements in saliva, the correlation between ferritin in saliva and serum, and the possible effects of correction of ferritin values by total protein were assessed. Ferritin was measured in saliva from healthy (n = 30) and COVID-19 (n = 65) patients with severe, (n = 18) or mild (n = 47) disease, depending on the need for nasal flow oxygen or assisted respiration. Ferritin was also measured in paired serum and saliva samples (n = 32) from healthy and COVID-19 patients. The evaluated inactivation protocols did not affect the assay’s results except the addition of 0.5% SDS. Significantly higher ferritin was found in the saliva of COVID-19 patients (median; 25–75th percentile) (27.75; 9.77–52.2 µg/L), compared with healthy controls (4.21; 2.6–8.08 µg/L). Individuals with severe COVID-19 showed higher ferritin values in saliva (48.7; 18.7–53.9) than mild ones (15.5; 5.28–41.3 µg/L). Significant correlation (r = 0.425; p < 0.001) was found between serum and saliva in ferritin. Ferritin levels were higher in COVID-19 patients in serum and saliva, and the highest values were found in those patients presenting severe symptomatology. In conclusion, ferritin in saliva has the potential to be a biomarker to evaluate severity in patients with COVID-19.

Objectives To evaluate four sample treatments in a safe and simple procedure for SARS-CoV-2 detection in saliva. Methods Four sample treatments in three-step procedure for the detection of SARS-CoV-2 in saliva, consisting in 1) heating at 95 °C during 5 minutes for sample inactivation, 2) sample treatment, and 3) analysis by RT-LAMP were evaluated using saliva samples with known amounts of added viral particles and saliva from infected individuals. Results Three treatments had a limit of detection of 500.000 viral particles per ml of saliva and could have a practical use for detecting those individuals that potentially could transmit the disease. The treatment consisting of a combination of phosphate buffer, dithiothreitol, ethylenediaminetetraacetic acid and proteinase K, and an additional 95 °C heating yielded the lower LOD 95 and the sensitivity ranged from 100% in patients with RT-PCRs NPS of Ct<20 to 47.8% in patients with Cts>30. Conclusions This report highlights the importance for an adequate sample treatment in saliva for the detection of SARS-CoV-2 and describes a cheap and flexible procedure that can be adapted to-point-of-care and, although its sensitivity is lower than RT-PCRs, can contribute to the Covid-19 control by the detection of individuals able to transmit the disease.