Abstract Background helping older people to maintain their independence, and identifying risk factors that compromise this, is of high importance. Polypharmacy is common in the very old (aged ≥ 85) but whether it can shape transitions in dependency in this fastest growing subpopulation is unclear. Methods using Newcastle 85+ Study data and multi-state modelling, we investigated the association between each additional medication prescribed and the progression of and recovery from dependency states, over 10 years (age 85–95). Participants were defined as either free from care (independent), requiring care less often than daily (low dependency), or requiring care at regular intervals each day or 24 hourly (medium/high dependency). Results each additional medication prescribed was associated with a 10% decreased chance of recovery from low dependence to independence (hazard ratio (HR): 0.90, 95% confidence interval (CI): 0.82–0.99). Discussion when a relatively able person visits the GP or clinical pharmacist, careful consideration should be given to whether the potential benefits from adding a new medication outweigh the risk to reduced recovery of independence.