How Chronic Is Polypharmacy in Old Age?

A Longitudinal Nationwide Cohort Study

Jonas W. Wastesson, PhD; Lucas Morin, MSc; Marie-Laure Laroche, PhD; Kristina Johnell, PhD


J Am Geriatr Soc. 2019;67(3):455-462. 

In This Article


Of 1,752,022 older adults (65 years or older) alive at baseline, 769,286 were exposed to polypharmacy. After excluding 57,854 individuals who died during the first 12 months of follow-up, the study population thus consisted of 711,432 older adults (Supplementary Figure S1). This represents 44% of the population aged 65 years or older in Sweden. Mean age at baseline was 77.4 years (SD = 7.8 years), and 59.1% were women. About 3% (n = 21,361) of study participants started a new episode of polypharmacy (ie, had not been exposed to polypharmacy during the 6-month period before baseline) (Table 1). Persons with a new episode of polypharmacy were, on average, younger, had fewer chronic conditions, and used fewer drugs at baseline (Table S1).

Polypharmacy was often long lasting. Overall, 82.3% of participants were exposed to polypharmacy for 6 months or longer, and 74.3% for 12 months or longer. Among older adults with a new polypharmacy episode, these proportions were 29.8% and 18.6%, respectively (Table 2). The proportion of individuals who remained exposed to polypharmacy until the end of follow-up was 55.3% in the total study population, but only 9.3% among people who had not been exposed to polypharmacy before baseline. Among the 317,478 older adults who discontinued polypharmacy, 76.3% experienced at least one more episode of polypharmacy during the follow-up period (Table S2). As shown in Figure 2, polypharmacy persisted for a longer time among older adults aged 75 years or older than among younger individuals. Episodes of polypharmacy were also longer among individuals with a higher number of medications at baseline (Figure S3).

Figure 2.

Kaplan-Meier survival functions. Solid-line curves denote the persistence of polypharmacy with a 2-month grace period. Dotted-line curves denote the persistence of polypharmacy with no grace period (sensitivity analysis). Percentages indicate polypharmacy exposure at 6 and 12 months. [Color figure can be viewed at]

Figure S3.

Kaplan-Meier survival functions. Duration of polypharmacy with a 2-month grace period, according to the number of drugs at baseline

During follow-up, we observed 21.2 million person-months with polypharmacy of a total of 25.3 million person-months. The average fraction of time with polypharmacy was thus 84%, ranging from 80% among individuals aged 65 to 74 years to 89% among those aged 95 years and older. Table 3 shows the proportion of older adults with a high fraction of time with polypharmacy (ie, exposed to polypharmacy for 80% or more of follow-up). In the total study population, 79.9% of older adults had a high fraction of time with polypharmacy, compared with 23.6% among persons with a new polypharmacy episode at baseline. After adjustment for potential confounders, this proportion increased with age, as well as with multidose drug dispensing compared with ordinary prescriptions (adjusted predicted probability, 93% vs 78%; P < .01). The proportion of nursing home residents with a high fraction of time with polypharmacy was higher than among community dwellers (90.7% vs 79.1%). However, after adjustment for other covariates, this association was reversed (predicted probability, 76.7% vs 80.1%). In post hoc analysis, we explored the interaction between living arrangement and drug dispensing scheme. This showed that community dwellers with multidose dispensing were, in fact, more likely to have a high fraction of time with polypharmacy than persons living in an institution (Table S3). In sensitivity analyses, the fraction of time with polypharmacy was calculated without the 1-month grace period, which yielded similar numbers, and a cutoff value of 50% or more was used, which left the association with other covariates largely unaffected, although a larger proportion of older adults were classified as chronic polypharmacy users (Table S4 and Table S5).