Epidemiology of Polypharmacy in the General Population: 27-Year Prospective Cohort Study

At least a third of the individuals with a noncommunicable disease (NCD), such as diabetes, chronic lung disease, or heart disease, are diagnosed with another NCD in their lifetime. This high burden of co-occurring chronic diseases (“multimorbidity”) necessitates frequent and simultaneous use of various drugs according to contemporary guidelines. Several studies reported the frequency and composition of polypharmacy (concurrent use of ≥5 drugs) in the general population, ^, but most were unable to capture the incidence and chronicity of polypharmacy because of their cross-sectional design. In this longitudinal study, we used data from the population-based Rotterdam Study to quantify the lifetime occurrence and chronicity of polypharmacy. We prospectively included 11,672 individuals aged ≥45 years that were free from polypharmacy at baseline. Continuous linkage of pharmacy dispensing records with the study database provides daily information on drug use. These records contain automated information on all dispensed prescriptions and, for example, include the Anatomical Therapeutic Chemical (ATC) codes, number of filled tablets and capsules or other dosage forms, dates of delivery, prescribed daily numbers or dosages, and prescription lengths. This means that this study only included prescription drugs and, for example, did not record nonprescription drugs or over-the-counter drugs. We considered an individual exposed to polypharmacy when the sum of simultaneously dispensed drugs on an ATC-4 level was ≥5 anywhere within a time frame of 90 days. We additionally expanded this definition to “hyperpolypharmacy” (≥10) and “excessive polypharmacy” (≥15 drugs). We calculated the remaining lifetime occurrence (“risk”) of polypharmacy with age as a time scale in left truncated data, while taking into account the occurrence of death as competing event. Follow-up started at study entry (with the age of 45 years as minimum) and ended at the yearly index date of polypharmacy, death, lost to follow-up, or the administrative study end date of January 1, 2018, whichever occurred first. Median age at baseline was 65.1 years (range 45-107 years), and 57.5% of the population were women. During the 27 years of follow-up (99.3% of potential person-years observed), 6755 individuals were exposed to polypharmacy, of whom 2035 were exposed to hyperpolypharmacy, and 320 to excessive polypharmacy. Lifetime occurrence of polypharmacy from the age of 45 years onward was 84.1% (95% CI 82.4-85.9) for men and 88.8% (95% CI 87.2-90.4) for women (P value for sex difference < 0.001; Figure 1). Approximately one-fourth of all individuals (24.7%, 95% CI 23.0-26.4, for men; 28.8%, 95% CI 27.2-30.4, for women; P value < .001) were concurrently exposed to ≥10 drugs at a certain point in their lifetime, whereas lifetime occurrence of excessive polypharmacy was 3.3% (95% CI 2.5-4.0) for men, and 4.2% (95% CI 3.5-4.9) for women (P = .047). Polypharmacy was often chronic, occurring in ≥2 consecutive years in two-thirds of individuals (69.1%). Lifetime occurrence of polypharmacy was high for individuals with and without major NCDs (cardiovascular disease, diabetes, lung disease, cancer, stroke, or neurodegenerative disease) at baseline (94.9%, 95% CI 91.4-98.5, and 87.6%, 95% CI 86.7-88.7, respectively). However, individuals free from NCDs at baseline were up to 13 to 19 years older when first exposed to polypharmacy than those with NCDs at baseline. This longitudinal overview on a population level provides insight into the frequency and chronicity of polypharmacy across a lifetime. Nine of 10 individuals aged 45 years and older are exposed to polypharmacy in their lifetime, and almost a quarter of people received ≥10 drugs by the age of 90 years. We observed small sex differences in the lifetime occurrence of polypharmacy, with an approximately 5% higher occurrence in women. In part, this may be explained by a longer life expectancy of women compared with men, but could also be attributed to sex differences in the type of drugs that underlie polypharmacy or to underlying indications for drug prescriptions. Compared with women, men for instance more often received drugs related to the prevention or treatment of prevailing cardiometabolic diseases that contributed to polypharmacy, such as cardiovascular medication (90% vs 85% for women) and drugs used in diabetes (20.1% vs 16.2% for women). In contrast, women more often received combinations of drugs (such as supplements, analgesics, hypnotics, and sedatives; 45.1% compared with 33.4% for men). This study extends prior evidence coming from registry studies among older adults (≥65 years) ^{,  ,}  and reveals that half (50.4%) of the population is already exposed to polypharmacy before the age of 65 years. Because only few randomized controlled trials enroll people with multiple chronic conditions, little is known about the efficacy and adverse effects of established drugs among older individuals with multimorbidity. Moreover, suboptimal combinations of different medications or drug-drug interactions are common in these individuals. Several initiatives have recently been put forward to reduce these potential harms, such as the STOPP/START or Beers criteria that assess inappropriate prescribing of medications. ^, An advantage of these qualitative criteria compared to a quantitative expression of polypharmacy is that the former more accurately captures the proportion of potentially inappropriate medication. It is thus possible that exposure to polypharmacy for some individuals in this study may well have been deemed appropriate according to these criteria. At the same time, it is considered equally important to evaluate the prescription of yet another drug, particularly in older, potentially frail individuals. Against that background, these findings can inform the development of new criteria and serve as the basis to facilitate clinicians and other stakeholders in differentiating between appropriate and inappropriate forms of polypharmacy.