This paper attempts a first approach to the discourses built around the pandemic unleashed in the year 2020. What issues are associated with the pandemic and COVID-19? Which are the most relevant? What is its temporal evolution in the first moments of it? With this objective, a source rarely used in social sciences will be used: the comments that readers of digital news produce in media forums. In turn, the work aims to provide elements that allow pondering the usefulness of natural language processing techniques to address this type of problem in social sciences. For the construction of the data, web scraping techniques were used and natural language processing algorithms were applied for its analysis. A relevant finding is linked to the apparent stability in the evolution over time of the topics, regardless of the metric used and the newspaper analyzed.