As people age, they are increasingly likely to suffer from more than one disease (multimorbidity): 60% of adults >65 suffer from more than three chronic diseases (1). These multimorbid patients may take different drugs to treat each disease (polypharmacy) (2), some of which may be inappropriate (3). Multimorbid patients are also at risk of underusing drugs (4). Both under- and overuse (inappropriate prescribing) may lead to avoidable hospital admissions and reduce the patient's quality of life (5–7). Therefore, polypharmacy among the multimorbid older patients must be managed carefully to meet the needs of the aging population. Multimorbidity, older age, and polypharmacy are associated with cardiovascular risk factors and disease (8): more than 50% of multimorbid patients have cardiovascular disease (CVD) (9) and are often prescribed statins to reduce cardiovascular events (CVE) and mortality in primary and secondary prevention (10, 11).

Statins are effective for secondary prevention of CVD even in old patients, with data until 82 years (12), but guidelines still suggest physicians consider individual factors that could influence their decision to start, change, or end statin therapy in older patients (13, 14). A small randomized controlled trial even showed that stopping statins in patients with life-limiting disease was safe, had no significant effect on mortality, and slightly increased quality of life (15, 16).

For primary prevention of atherosclerotic CVD, the current European Society of Cardiology and American Heart Association (AHA) guidelines recommend statins to match individual cardiovascular risk, setting the minimum bar at ≥5 to <10% 10-year risk of fatal and non-fatal atherosclerotic CVD (ASCVD) (17–19). Cardiovascular risk can be calculated with tools like the PROCAM-SCORE or SCORE2 (20, 21), which include age, gender, and known modifiable cardiovascular risk factors. Among older adults, the evidence of benefits of statins for primary prevention is less clear. The recent ESC guidelines for elderly recommend risk-factor treatment in older patients (≥70 years) with very high-risk for CVD (≥15% 10-year risk of CVD according to SCORE-OP2,) and apparently healthy older high-risk patients (7.5%–15% 10-year CVD risk, Class IIa recommendation) but there is insufficient evidence for low-density lipoprotein cholesterol (LDL-C) targets (22). The AHA guidelines neither suggest nor discourage statin prescription as primary prevention for this group of patients and they encourage physicians to individualize their recommendations, taking into account life expectancy, patient preferences, co-morbidities, and other factors (evidence level E) (13, 18, 19).

Based on current evidence, physicians should carefully consider whether to prescribe drug treatment in multimorbid older patients with polypharmacy since data have shown that such patients are at risk for inappropriate prescription (underuse and overuse) (3). To date, there is little data on whether statins are appropriately prescribed to multimorbid older patients with polypharmacy, and risk factors associated with statin under- and over prescription.

To overcome this research gap, we conducted a cross-sectional analysis of the OPERAM trial (OPtimising thERapy to prevent Avoidable hospital admissions in Multimorbid older people) to assess the prevalence and determinants of potentially inappropriate statin therapy in multimorbid older hospitalized patients.

We included participants from the Swiss study center (Bern University Hospital) of the cluster-randomized controlled OPERAM trial with detailed information on statin use, related factors and lipid values. A detailed protocol of OPERAM and its main results were published elsewhere (3, 23). We included only Swiss participants as the adjudication algorithm was based on local guidelines. Since the OPERAM trial studied a systematic drug review vs. usual care, we chose a cross-sectional study design at hospital admission, prior to the drug review, in order to receive the most accurate real life reflection.

We considered all patients with a diagnosis of cardiovascular ischemic disease to be patients who needed secondary prevention for cardiovascular events. Thus, all patients with cardiovascular disease were considered to be underusing a statin if they were not taking a statin. For all patients, we calculated the Walter Score to determine individual 1-year mortality risk (28); we considered a score >6 (>60% 1-year mortality) to be potentially life limiting. Since current evidence suggests that statin therapy offers no benefit when prescribed for primary prevention to patients with a life-limiting disease, we classed all patients whose Walter Score was >6 as potentially being overusing statins if they were using statins at baseline (Figure 1) (15, 29).

We calculated individual risk for cardiovascular events within the next 10 years based on the patient's AGLA Score (PROCAM Score adapted for Switzerland) (25). Because several guidelines make no clear recommendation for primary prevention in patients aged >75 years (18, 19, 22, 30), we calculated cardiovascular risk scores for those patients as if they were 75 years old. Additionally, the Walter Score was used to estimate individual life expectancy (above) and two medical doctors (LA, OB) independently reviewed the patients' chart to determine if they were on appropriate statin therapy. If the two reviewers disagreed, they consulted a third reviewer (MF). Figure 1 shows the study flow chart for adjudication, which we developed based on current guidelines.

The following predictors were included into our analysis: age (classed into three categories: 70–75, 76–85 and >85 years), sex, being housebound [defined as inability to leave the house unassisted e.g., for (primary) care visits], ≥5–9 chronic medications vs. ≥10 chronic medicationsco-morbidities (into ≥10 and <10 chronic diagnoses), and whether they suffered from dementia/cognitive impairment according to ICD-10 code. All these factors potentially influence the likelihood medication will be prescribed inappropriately (31).

The Swiss OPERAM study center included 822 hospitalized multimorbid patients ≥70 years with polypharmacy (≥5 medications); 17 patients (2.1%) withdrew their consent; and 90 patients (10.9%) were excluded because their available information did not allow us to adjudicate the appropriateness of statin therapy. We thus included 715 patients in our analysis (see the study flow chart in the Supplementary Figure S1). Included patients had a mean age of 79.7 years [±6.5 years standard deviation (SD)]; 60% were male (Table 1). In the secondary prevention group (n = 437), 295 patients (67.5%) used a statin and 105 (24%) patients had a Walter score >6, 68 (23%) of those using a statin (Supplementary Table S3). In the primary prevention group (n = 278), 82 patients (29.5%) used statins and 89 (32%) patients had a Walter Score >6.

Additional baseline characteristics of included patients are presented in Table 1 and comparisons of the baseline characteristics of patients we included and patients we excluded because we could not properly adjudicate them in Supplementary Table S1.

In the 278 patients in primary prevention, we determined estimated risk for CVE at 10-years was low (<10%) in 95 patients, intermediate (10%–20%) in 53 patients, and high (>20%) in 40 patients (Supplementary Table S3). Patients with a Walter Score >6 (n = 89) were classified as potential Statin-overusers without further risk calculation.

Appropriateness of statin therapy in secondary cardiovascular prevention:

Of the 437 patients in secondary prevention, 274 (62.7%) were using a statin appropriately, 105 (24%) patients were underusing a statin, while 58 (13.3%) patients were considered to be potentially overusing a statin as their predicted mortality risk was >60% based on their Walter Score (≥6 points) (Table 2).

In secondary prevention, statin underuse was significantly associated with being very old (>85 years; OR 2.40; 95% CI 1.26–4.57) and female (OR 2.79; 95% CI1.78–4.38) in bivariate models (Table 3). We confirmed the association between underuse and being female with LASSO regression (OR 2.65; 95%CI 1.67–4.22).

Potential overuse was associated with being housebound (OR 3.58; 95% CI 1.38–9.28) and taking ≥10 medications (1.94; 95%CI 1.05–3.57) in bivariate analyses, which was confirmed after LASSO regression (OR 3.53 (1.32–9.46) and 1.95 (1.05–3.67)) (Table 4).

Among 715 multimorbid older (≥70 years) hospitalized patients, potentially inappropriate statin therapy was common (33.7%) due to potential underuse with (15.5%) and potential overuse (18.2%). Female gender was significantly associated with potential underuse in secondary prevention (OR 2.65). Being housebound (OR 3.53) and taking ≥10 medications (OR 1.95) were associated with potential overuse in secondary prevention. No specific risk factors were identified for potential over- or underuse in primary prevention.

Though the benefits of statins for primary prevention in the older patients remain controversial, statin therapy in secondary prevention in older patients is supported by evidence (22). Despite the evidence, this cross-sectional study found that 24% of patients in secondary prevention were potentially underusing statins. In secondary prevention, women, particularly, were at risk of underuse (OR 2.65), a finding that aligns with previous studies (32–34).

Statin use is clearly established for prevention of further cardiovascular events in secondary prevention, in both men and women (35). On average, women live longer than men and are therefore more likely to suffer from CVE (36). CVD is the main cause of death for women in developed countries (37, 38), but health care systems may not adequately consider women's cardiovascular risk (38). Though men and women have the same main cardiovascular risk factors, women may present differently in clinical exams, have different symptoms, and tests may not work as well for women as men (39, 40). Our findings align well with a population-based Italian study of older patients with polypharmacy, in which women were more likely to discontinue statin therapy (41).

Housebound patients in secondary prevention were at risk for overuse (defined as having a Walter Score >6) of a statin (OR 3.53), a finding that aligns with a Japanese study of eldery, housebound patient amongst whom potentially inappropriate prescribing was highly prevalent (42). This might be due to the fact that preventive medical visits are associated with a high effort by these patients and his/her proxies (42). A further factor could be that stopping an established statin therapy is currently a matter of discussion and doctors might be especially hesitant to stop a statin in a patient with established ASCVD despite advanced chronic illness and a high 1-year mortality risk (Walter Score > 6) (16). However, Kutner et al. found that stopping statins in patients with limited life expectancy was safe and slightly improved quality of life (15). In primary prevention, we found no pre-defined risk factor associated with higher risk of inappropriate statin therapy (neither potential underuse nor potential overuse), perhaps because it was more difficult to assess cardiovascular risk and define the necessity for statin therapy in the primary prevention group. Though older people are at higher cardiovascular risk, few RCTs have studied the potential benefits of statin therapy for primary prevention in the elderly. A recent meta-analysis of 28 RCTs that analyzed the benefits of statin therapy in 186,854 individuals (but only 8% aged >70 years) found that statin therapy benefited to patients over 70, but provided little evidence on patients in primary prevention (43). An age-stratified analysis of the JUPITER-trial and HOPE III study showed that taking Rosuvastatin would benefit patients >70 years regarding a composite outcome (pooled estimate HR 0.74) of nonfatal MI, stroke or cardiovascular death (44) but included only few patients >80 years. On the other hand, deprescribing statins in older patients is common in primary prevention in the very older people because clinicians might weigh the potential of side effects in frail patients higher than the potential long-term benefit (45). In an international survey using case vignettes, GPs recommended to stop statins for primary prevention especially in frail patients, those with evident side effects and patients with limited life expectancy (16). In general, evidence towards statin therapy in older people in primary prevention is limited, as shown by the discrepancies in different guidelines concerning statin therapy in the elderly (46). Especially for multimorbid older patients guidelines or recommendations concerning (deprescribing) statins are lacking (47).

Our finding of ASCVD prevalence (61%) aligns with previous studies of this older multimorbid population (8). We found statin therapy was prescribed to 52.5% in the entire patient cohort, (29.5% of patients in primary prevention and 67.5% of patients in secondary prevention). A 2011–2015 UK cohort study of statin prescription in old patients (>80 years) in primary prevention yielded similar results (30%) (45). Other studies from the UK and the US found that between 63% and 80% of the older population are taking statins (45, 32).

A third of multimorbid older patients with polypharmacy either potentially overused or underused statins. Among participants in secondary cardiovascular prevention, women were at highest risk for potential underuse while housebound patients were at risk for potential overuse of a statin. Physicians should carefully evaluate their need for statin prescriptions in this understudied population and adapt prescriptions as needed.