Web tools are widely used among population to obtain health related information, and these data are often employed for public health monitoring. Here we analyzed searches related to smell loss and taste loss, recently linked to COVID-19, as well as sight loss and hearing loss, not included in the COVID-19 symptoms list. Google Trends results per region (Italy) or state (United States) over several weeks were compared to the number of new cases prevalence in that geographical area. Taste and smell loss searches were correlated with each other, and, during a limited time window, with new COVID-19 cases. However, this correlation decreased with time, attributable, at least in part, to media coverage. As new symptoms are being discovered for COVID-19 and the pandemic continues to spread around the globe, the lesson learned here, that correlation between public interest in novel symptoms of infectious disease has an initial spike (the "surprise rise") and subsequently goes to a new baseline due to "knowledge saturation", is of general and practical value for the public.Supplementary Figure S1 repeats the calculation presented in Figure 1 for the 22-28 April week. This clearly illustrates the current lack of correlation between the number of new cases in each region (Italy) or state (US) and the popularity of Google searches on either taste or smell loss.