Background
Opioids have been associated with an increased excess mortality-risk in the general population. Older people with dementia have been prescribed opioids more frequent and the use of strong opioids has increased considerably over the past decade among older people with dementia. Whether opioids are associated with an excess mortality risk in older people with dementia is unknown.

Method
Matched cohort study using nationwide registries. All Danish residents aged 65 years and older diagnosed with dementia between 01.01.2008-31.12.2018 were included. Exposure was defined as first opioid prescription after dementia diagnosis (1 year washout before diagnosis). Persons exposed to opioids were matched with up to two unexposed persons on age and sex. Cox proportional hazards models were used to compare rates of death within 180 days after the initiation of opioid treatment and adjusted for potential confounders (time since diagnosis, nursing home residency, comorbidity, admission within 30 days, number of medications).

Result
Forty-two percent (31,619/75,471) of older people with dementia redeemed a prescription for an opioid after their diagnosis of dementia. We matched 31,619 exposed persons to 63,235 unexposed. Among the exposed, 10,642 (33.7%) died within 180 days after initiating their first opioid prescription compared with 3980 (6.4%) of the unexposed, yielding a 4-fold increased excess mortality risk (adjusted Hazard ratio (95% CI): 4.13 (3.98-4.30)). Strong opioids were associated with a 6-fold increased mortality risk (6.34 (6.00-6.69)), whereas the risk was lower for weak opioids (2.52 (2.38-2.67). Among those who had redeemed transdermal fentanyl as their first prescription, 65.3% died within the first 180 days compared with 6.7% in the unexposed with a risk of 8.03 (7.02-9.18). The risk was greatest within the first 14 days, where the mortality (for all opioids) was increased 11-fold (10.95 (9.87-12.15)), but there was still a 2-fold increased mortality risk after 90 days (2.36 (2.21-2.52)).

Conclusion
New opioid use was frequent among elderly with dementia, and this was associated with a markedly increased excess mortality, which is worrisome. The data calls for more research to identify sub-groups with highest excess mortality risk to make a risk–benefit assessment and guidelines to assure appropriate prescribing.