Abstract Importance treatment of dementia in individuals with comorbidities is complex, leading to potentially inappropriate prescribing (PIP). The impact of PIP in this population is unknown. Objective to estimate the rate of PIP and its effect on adverse health outcomes (AHO). Design retrospective cohort. Setting primary care electronic health records linked to hospital discharge data from England. Subjects 11,175 individuals with dementia aged over 65 years in 2016 and 43,463 age- and sex-matched controls. Methods Screening Tool of Older Persons’ Prescriptions V2 defined PIP. Logistic regression tested associations with comorbidities at baseline, and survival analyses risk of incident AHO, adjusted for age, gender, deprivation and 14 comorbidities. Results the dementia group had increased risk of PIP (73% prevalence; odds ratio [OR]: 1.92; confidence interval [CI]: 83–103%; P < 0.01) after adjusting for comorbidities. Most frequent PIP criteria were related to anti-cholinergic drugs and therapeutic duplication. Risk of PIP was higher in patients also diagnosed with coronary-heart disease (odds OR: 2.17; CI: 1.91–2.46; P < 0.01), severe mental illness (OR: 2.09; CI: 1.62–2.70; P < 0.01); and depression (OR: 1.81; CI: 1.62–2.01; P < 0.01). During follow-up (1 year), PIP was associated with increased all-cause mortality (hazard ratio: 1.14; CI: 1.02–1.26; P < 0.02), skin ulcer and pressure sores (hazard ratio: 1.66; CI: 1.12–2.46; P < 0.01), falls (hazard ratio: 1.37; CI: 1.15–1.63; P < 0.01), anaemia (hazard ratio: 1.61; CI: 1.10–2.38; P < 0.02) and osteoporosis (hazard ratio: 1.62; CI: 1.02–2.57; P < 0.04). Conclusion patients with dementia frequently receive PIPs, and those who do are more likely to experience AHO. These results highlight the need to optimise medication in dementia patients, especially those with comorbidities.