Mortality, hospitalisation, institutionalisation in community-dwelling oldest old: The impact of medication

• In a primary care sample of oldest old (80+), the mean chronic medication use was 5.4.
• High medication use was strongly associated with hospitalisation (11% increase per extra medication).
• High medication use was less clearly associated with increased mortality.
• High medication use was not associated with entering a nursing home.

BackgroundHigh drug use and associated adverse outcomes are common in older adults. This study investigates association of medication use with mortality, hospitalisation, and institutionalisation in a cohort of community-dwelling oldest old (aged 80 and over).MethodsBaseline data included socio-demographic, clinical, and functional characteristics, and prescribed medications. Medications were coded by the Anatomic Therapeutic Chemical classification. Survival analysis was performed at 18 months after inclusion using Kaplan–Meier, and multivariate analysis with Cox regression to control for covariates.ResultsPatients’ (n = 503) mean age was 84.4 years (range 80–102), and 61.2% was female. The median medication use was 5 (0–16). The mortality, hospitalisation, and institutionalisation rate were 8.9%, 31.0%, and 6.4% respectively. The mortality and hospitalisation group had a higher level of multimorbidity and weaker functional profile. Adjusted multivariate models showed an 11% increased hospitalisation rate for every additional medication taken. No association was found between high medication use and mortality, nor with institutionalisation. A higher association for mortality was observed among verapamil/diltiazem users, hospitalisation was higher among users of verapamil/diltiazem, loop diuretics and respiratory agents. Institutionalisation was higher among benzodiazepines users.ConclusionIn the community-dwelling oldest old (aged 80 and over), high medication was clearly associated with hospitalisation, independent of multimorbidity. The association with mortality was clear in univariate, but not in multivariate analysis. No association with institutionalisation was found. The appropriateness of the high medication use should be further studied in relation to mortality, hospitalisation, and institutionalisation for this specific age group.