Abstract Background The Coronavirus disease 2019 (COVID-19) pandemic is an unprecedented traumatic event influencing the healthcare, economic, and social welfare systems worldwide. In order to slow the infection rates, lockdown has been implemented almost everywhere. Italy, one of the countries most severely affected, entered the “lockdown” on March 8, 2020. Methods The COvid Mental hEalth Trial (COMET) network includes 10 Italian university sites and the National Institute of Health. The whole study has three different phases. The first phase includes an online survey conducted between March and May 2020 in the Italian population. Recruitment took place through email invitation letters, social media, mailing lists of universities, national medical associations, and associations of stakeholders (e.g., associations of users/carers). In order to evaluate the impact of lockdown on depressive, anxiety and stress symptoms, multivariate linear regression models were performed, weighted for the propensity score. Results The final sample consisted of 20,720 participants. Among them, 12.4% of respondents ( N = 2,555) reported severe or extremely severe levels of depressive symptoms, 17.6% ( N = 3,627) of anxiety symptoms and 41.6% ( N = 8,619) reported to feel at least moderately stressed by the situation at the DASS-21. According to the multivariate regression models, the depressive, anxiety and stress symptoms significantly worsened from the week April 9–15 to the week April 30 to May 4 ( p < 0.0001). Moreover, female respondents and people with pre-existing mental health problems were at higher risk of developing severe depression and anxiety symptoms ( p < 0.0001). Conclusions Although physical isolation and lockdown represent essential public health measures for containing the spread of the COVID-19 pandemic, they are a serious threat for mental health and well-being of the general population. As an integral part of COVID-19 response, mental health needs should be addressed.