Objective: To identify predictors of persistent cognitive impairment at 12 months after hospitalization due to COVID-19 (SARS-CoV-2) infection. Design: Retrospective, single-centre study. Subjects: All consecutive patients assessed in physical and rehabilitation medicine consultations at 3 months with a neuropsychiatric testing (NPT) at 6 months. Methods: A Mini Mental State Examination (MMSE) was performed at 3 months and NPT at 6 and 12 months, exploring global cognitive efficiency, attention and processing speed, short-term memory and executive function. Logistic regression and receiver operating characteristic curves were used to identify predictors of persistent cognitive impairment. Results: Among 56 patients, 64.3% and 53.6% had 1 or more impaired cognitive functions at 6 and 12 months, respectively, attention and processing speed being the most represented (41.1% at 12 month). Duration of oxygen therapy (odds ratio 0.926 [0.871–0.985], p = 0.015) and MMSE score at 3 months (odds ratio 0.464 [0.276–0.783], p = 0.004) were associated with cognitive impairment at 12 months by multivariable analysis (R² 0.372–0.497). Conclusions: Half of patients have cognitive impairment 12 months after acute SARS-CoV-2 infection requiring hospitalization. The duration of oxygen therapy in acute care could be a protective parameter. Systematic evaluation with the MMSE at 3 months after infection might be an effective tool to detect risk.