Assessment of saliva interference with UV-based disinfection technologies

Worldwide shortages of personal protective equipment during COVID-19 pandemic has forced the implementation of methods for decontaminating face piece respirators such as N95 respirators. The use of UV irradiation to reduce bioburden of used respirators attracts attention, making proper testing protocols of uttermost importance. Currently artificial saliva is used but its comparison to human saliva from the UV disinfection perspective is lacking. Here we characterize UV spectra of human and artificial saliva, both fresh and after settling, to test for possible interference for UV-based disinfection. ASTM 2720 artificial saliva recipe (with either porcine or bovine mucin) showed many discrepancies from average (N = 18) human saliva, with different mucins demonstrating very different UV absorbance spectra, resulting in very different UV transmittance at different wavelength. Reducing porcine mucin concentration from 3 to 1.7 g/L brought UVA254 in the artificial saliva to that of average human saliva (although not for other wavelengths), allowing 254 nm disinfection experiments. Phosphate saline and modified artificial saliva were spiked with 8.6 log CFU/ml B. subtilis spores (ATCC 6633) and irradiated at dose of up to 100 mJ/cm2, resulting in 5.9 log inactivation for a saline suspension, and 2.8 and 1.1 log inactivation for ASTM-no mucin and ASTM-1.7 g/L porcine mucin 2 ILL dried droplets, respectively. UVC irradiation of spores dried in human saliva resulted in 2.3 and 1.5 log inactivation, depending on the size of the droplets (2 vs 10 ILL, respectively) dried on a glass surface. Our results suggest that in the presence of the current standard dried artificial saliva it is unlikely that UVC can achieve 6 log inactivation of B. subtilis spores using a realistic UV dose (e.g. less than 2 J/cm2) and the ATSM saliva recipe should be revised for UV decontamination studies.