Reply to Kao and Liaw

Abstract

The letter by Kao and Liaw raises some interesting observations. However, none ofthem question the results and conclusions of our study. As a first point, theyrecommend the use of fecal specimens to avoid, in their words, an uncertainty inthe accuracy of RT-PCR from respiratory samples. While it has been reported thatfecal samples might provide increased sensibility to detect SARS-CoV-2 at laterstages of infection, no evidence supports that data collected in this way were moreaccurate or relevant in terms of diagnosis or transmission.Secondly, Kao and Liaw raise the issue of the time elapsed between symptom onsetand RT-PCR test. This is a well-known confounder in viral load comparisons, andas such, it was clearly acknowledged in our study. We showed that the periodbetween symptom onset and diagnosis were not significantly different among thegroups studied. In fact, there were no cases with more than 7 days between theRT-PCR test and symptom onset in the population under 10 years old, and theproportion was negligible for the older population.Finally, we share the interest of Kao and Liaw about the possible association ofrespiratory viral loads with the expression of host genes related to mucosalimmunity. The search for these associations was not in the scope of our study, butwe agree they could be informative about the mechanisms that control respiratoryviral loads.All in all, by studying the SARS-CoV-2 viral loads of 45,318 respiratory samplesfrom COVID-19 patients measured by a validated RT-PCR method, our studyclearly demonstrates that infants younger than 6 months have higher viral loadsthan any other age group.