Background The COVID-19 pandemic has raised concerns regarding its psychological effect on university students, especially healthcare students. We aimed at assessing the risk of mental health problems according to the type of university studies, by adjusting for potential confounders. Methods We used data from the COSAMe study, a national cross-sectional survey including 69,054 French university students during the first quarantine. The mental health outcomes evaluated were suicidal thoughts, severe self-reported distress (as assessed by the Impact of Events Scale-Revised), stress (Perceived Stress Scale), anxiety (State-Trait Anxiety Inventory, State subscale), and depression (Beck Depression Inventory). Multivariable logistic regression analyzes were performed to test the association between the type of university studies (healthcare studies: medical and non-medical, and non-healthcare studies) and poor mental health outcomes, adjusted for sociodemographic characteristics, precariousness indicators, health-related data, quality of social relationships, and data about media consumption. Results Compared to non-healthcare students (N = 59,404), non-medical healthcare (N = 5,431) and medical students (N = 4,193) showed a lower risk of presenting at least one poor mental health outcome (adjusted OR [95%CI] = 0.86[0.81-0.92] and 0.87[0.81-0.93], respectively). Compared to non-healthcare students, medical students were at lower risk of suicidal thoughts (0.83[0.74-0.93]), severe self-reported distress (0.75[0.69-0.82]) and depression (0.83[0.75-0.92]). Non-medical healthcare students were at lower risk of severe selfreported distress (0.79[0.73-0.85]), stress (0.92[0.85-0.98]), depression (0.83[0.76-0.91]), and anxiety (0.86[0.80-0.92]). Limitations This is a large but not representative cross-sectional study, limited to the first confinement. Conclusions Being a healthcare student is a protective factor for mental health problems among confined students. Mediating factors still need to be explored.