Objective: Studies that investigate predictive factors for spontaneous recovery (reversion) from mild cognitive impairment (MCI) are only beginning to emerge, and the long-term course of MCI is not properly understood. We aimed to investigate stability of the MCI diagnosis, predictors for reversion, as well as the trajectory of MCI over the course of 12 years. Materials and Methods: Data were drawn from the Swedish population study: Good Aging in Skåne with MCI defined according to the expanded Mayo Clinic criteria. A total of 331 participants, aged 60-95 years with MCI, were used to investigate 6-year MCI stability and reversion, and 410 participants were used to inspect 12-year MCI trajectory. Predictors for reversion included demographical factors, psychological status, and factors tied to the cognitive testing session and the operationalization of the MCI criteria. Results: Over half (58%, 95% CI 52.7-63.3) of the participants reverted back to normal cognitive functioning at 6-year follow-up. Of those with stable MCI, 56.5% (95% CI 48.2-64.8) changed subtype. A total of 23.9% (95% CI 13.7-34.1) of the 6-year follow-up reverters re-transitioned back to MCI at 12-year followup. ORs for reversion were significantly higher in participants with lower age (60-year-olds: OR 2.19, 95% CI 1.08-4.43, 70-year-olds: OR 3.11, 95% CI 1.27-7.62), better global cognitive functioning (OR 1.15, 95% CI 1.03-1.29), good concentration (OR 2.53, 95% CI 1.06-6.05), and single-domain subtype (OR 2.68, 95% CI 1.51-4.75). Conclusion: Our findings provide further support that MCI reversion to normal cognitive functioning as well as re-transitioning to MCI is fairly common, suggesting that the MCI trajectory does not necessarily lead straight to dementia. Additionally, assessment of factors associated with reversion can aid clinicians to make accurate MCI progression prognosis. impaired daily functional abilities or dementia. Of the excluded 1,188, 551 participants selected at random were excluded for the purpose of creating a normative sample. The study was approved by the regional Ethics Committee of Lund University, and written consent from all participants was obtained.The participant was examined by a nurse, a physician, and psychological test administrator. Participants filled out self-rated questionnaires regarding level of education, marital status, and functional ability. A cognitive assessment including tests measuring 4 cognitive domains: memory (3 tests: 16-item word-recall