Introduction SARS-CoV-2 transmission within schools and its contribution to community transmission are still a matter of debate. Methods A retrospective cohort study in all public schools in Catalonia was conducted using publicly available data assessing the association between the number of reported SARS-CoV-2 cases among students and staff in weeks 1–2 (Sept 14-27th, 2020) of the academic year with school SARS-CoV-2 incidence among students in weeks 4–5. A multilevel Poisson regression model adjusted for the community incidence in the corresponding basic health area (BHA) and the type of school (primary or secondary), with random effects at the sanitary region and BHA levels, was performed. Results A total of 2184 public schools opened on September 14th with 778,715 students. Multivariate analysis showed a significant association between the total number of SARS-CoV-2 cases in a centre in weeks 1–2 and the SARS-CoV-2 school incidence among students in weeks 4–5 (Risk Ratio (RR) 1.074, 95% CI 1.044–1.105, p-value <0.001). The adjusted BHA incidence in the first two weeks was associated with school incidence in weeks 4–5 (RR 1.002, 95% CI 1.002–1.003, p-value <0.001). Secondary schools showed an increased incidence in weeks 4 and 5 (RR primary vs secondary 1.709 95% CI 1.599–1.897, p-value <0.001). Conclusions Safety measures adopted by schools were not enough to stop related-to-school transmission in students and could be improved. The safest way to keep schools open is to reduce community transmission down to a minimum.