Our cross-sectional study revealed a high prevalence of multimorbidity and polypharmacy in the elderly aged 65 years or older, which is consistent with previous studies [15–18]. In most studies, the prevalence of multimorbidity increased with age, and the prevalence in older persons was nearly 100%. The prevalence of multimorbidity was 96.74% in our study, with 95.24% in the 65–74 age group, 98.75% in the 75–84 age group and 100% in the ≥ 85 age group. Our previous study with 46,799 hospitalized patients aged 60 years and older also revealed that up to 91.07% of patients had multimorbidity [15]. The prevalence of polypharmacy varies depending on age group, healthcare setting and region. In our study, 91.30% of patients had polypharmacy, and 43.48% took 10 or more medications, which is defined as hyperpolypharmacy, similar to a retrospective, cross-sectional study including 1200 inpatients aged ≥ 65 years in the geriatric department of the hospital, with a 91.58% prevalence of polypharmacy [18]. In a narrative review including systematic reviews and original studies published between 2015 and 2020 in the MEDLINE database, the prevalence of polypharmacy ranged from 4% among community-dwelling older people to over 96.5% among hospitalized patients [10].
We further evaluated PIMs based on both the Beers criteria and Chinese criteria, and similar results were identified for 66.30% (183/276) of the study population based on the Beers criteria and 55.07% (152/276) based on Chinese criteria. The prevalence of PIMs varies widely among different studies. In a systematic review and meta-analysis of 94 studies involving nearly 371.2 million people from 17 countries, the prevalence of PIMs ranged from 1.3–95.2%, with a 36.7% pooled prevalence of PIMs [19]. In their study, Africa had the highest prevalence of PIMs (47.0%; 95% CI, 34.7%-59.4%), followed by South America (46.9%; 95% CI, 35.1%-58.9%), Asia (37.2%; 95% CI, 32.4%-42.2%), Europe (35.0%; 95% CI, 28.5%-41.8%), North America (29.0%; 95% CI, 22.1%-36.3%), and Oceania (23.6%; 95% CI, 18.8%-28.8%). In a retrospective study conducted in a tertiary care centre in the USA, PIMs reached to 87.4% by the 2015 Beers criteria [20]. In a population-based cohort study including 1 105 295 individuals, 48.3% were prescribed at least one PIMs [21].
We analysed the specific subtypes of patients with PIMs and found most patients had 1 PIMs, 40.22% of patients had 1 PIMs based on Beers criteria, accounting for 60.66% of patients with PIMs. 41.30% of patients had 1 PIMs based on Chinese criteria, accounting for 75% of patients with PIMs.
In order to investigate possible risk factors influencing the incidence of PIMs, we compared patients with different number of PIMs, and found the number of prescribed medications in patients with PIMs was significantly higher than non-PIMs group, patients with ≥ 2 PIMs had more prescribed medications than patients with 1 PIMs analysed by Mann–Whitney U test. Similarly, Chi-square test indicated that the incidence of PIMs increased with the number of prescribed medications. Further analysis by logistic regression also showed the strongest predictor for PIMs by both critera was number of prescribed medications. Compared to prescriptions of 1–4 medications, the risk of PIMs for prescriptions of 5–9 medications and 10 or more medications increased to 5.161-fold and 15.858-fold. Several studies have explored possible factors related to PIMs. A retrospective study conducted in Brazil including 314 patients aged ≥ 60 years showed that factors associated with PIMs were the number of medications, length of hospital stay and excessive potential drug interactions [22]. In a cross-sectional study in Saudi Arabia including 23,417 people aged ≥ 65 years over three years (2017–2019), factors associated with PIMs involved ≥ 5 dispensed medications and ≥ 5 diagnoses [23]. Female gender and increasing number of medications were the independent predictors of PIMs among older patients discharged from Tawam hospital according to the Beers criteria 2019 [24].
As mentioned in above studies and our study, an increased number of prescribed medications is a common risk factor associated with PIMs. Medication reviews and interventions by combination of clinical physicians and pharmacists without affecting efficacy is mentioned in many articles as an effective strategy for minimizing the incidence of polypharmacy and PIMs [3, 25–28]. Deprescribing is the medically supervised process of decreasing or stopping medications that are no longer needed or may cause harm [3]. By deprescribing, it may decrease medication side effects and drug-drug interactions, reduce non-adherence to medications, improve health and lower costs, thus patients may benefit a lot. In a systematic review and meta-analysis, medication deprescribing interventions significantly lowered PIMs in 7 studies [25]. In a randomized controlled trial, drug-intervention may improve drug appropriateness in community-dwelling polymedicated elderly people [26]. Rodrigues et al reported in a systematic review including 47 articles medication review was the most successful intervention among various types of interventions in hospitals [27]. In another study, clinical pharmacist recommendations reduced the total number of medications by 7.5% from 13.4 to 12.4 per patient, the total number of prescribed PIMs by 21.8% from 312 to 244 for psychogeriatric primary care patients [28].
The most common PIMs identified by the Beers criteria and Chinese criteria in our study were proton pump inhibitors (PPIs) and clopidogrel. PPIs were reported to be the most common PIMs based on the Beers criteria, and clopidogrel was reported to be a common PIMs according to Chinese criteria in several studies[21, 29–30]. Benzodiazepines were very common PIMs detected by both the Beers and Chinese criteria in our study, which was consistent with previous reports [19, 21, 31]. In our follow-up, we found that benzodiazepines had the highest incidence of ADRs, with drowsiness as the most common side effect.
Limitations should be noted in our study. First, this was a cross-sectional, observational and single-center study in which only inpatients were included; more elderly individuals, including community-dwelling individuals and those in nursing homes, should be included. In addition, comparison of the results of medication reviews and interventions should also be included in further studies.