Association of potentially inappropriate medications with outcomes of inpatient geriatric rehabilitation

The proportion of older people in western countries is increasing. Higher age is associated with multimorbidity, which is a widespread medical problem often leading to polypharmacy. Polypharmacy is associated with an increased risk of prescription of potentially inappropriate medications (PIM). This study investigated whether PIM have an association with rehabilitation outcomes of an inpatient geriatric rehabilitation in Switzerland. The use of PIM is often found in a population of older persons [7, 16], and may include medications that are not indicated, which are not correctly prescribed, or which are not appropriate for elderly people. In an evaluation of medication use over a period of 7 days, Fialová et al. showed that over 95% of elderly persons took at least one medication and polypharmacy was documented in 51% of the patients [7]. At least one potentially inappropriate drug was taken by 19.8% of the patients. Frequently prescribed PIMs were vasodilators, benzodiazepines, heart medication, antidepressants and anticoagulants (e. g. warfarin) and PIMs were associated with an age of 75–79 years, frailty and polypharmacy [21]. Frail people are at increased risk of adverse drug events, e. g. falls, hospitalization and mortality. Landi et al. have shown that PIM may have adverse effects, which decrease the physical performance of patients, especially in community-dwelling elderly people over 80 years of age [13]. The physical performance decreased more in patients taking two or more inappropriate drugs in comparison with non-users or patients taking only one inappropriate drug, indicating that the association between impaired physical performance and PIM becomes stronger when the number of inappropriate drugs increases. Users of PIM were more likely to have cognitive impairment in comparison with non-users, were less likely to be physically active, and had a higher number of medications.

To identify PIM most of the previous studies used the criteria of Beers [4], a list of medications that should be avoided in elderly patients due to frequent adverse drug events; however, the drugs on the Beers list are mostly administered in the USA and are not available in Europe. Alternatives in the European setting include the Priscus list [11], a list of medications administered in Germany that should be avoided, the Screening Tool to Alert doctors to the Right Treatment (START) or the Screening Tool of Older Persons’ potentially inappropriate Prescription (STOPP), which focus on avoiding inappropriate drugs (STOPP) or to identify undertreatment medications (START) [10]. Gallagher and O’Mahoney showed that the STOPP criteria identified a significantly higher proportion of patients requiring hospitalization as a result of PIM-related adverse events than Beers’ criteria [9].

Rehabilitation is known to have a beneficial effect on the physical performance of elderly patients [15]. Baztán et al. showed that patients cared for in acute geriatric rehabilitation units were more likely to live at home after discharge, and that the effects were maintained 3 months after discharge [3]. In another meta-analysis Bachmann et al. found that inpatient rehabilitation specifically designed for geriatric patients has the potential to improve outcomes regarding function (Barthel index), nursing home admissions and mortality [1]. Inpatient rehabilitation was significantly more beneficial regarding functional outcomes at discharge and in the long term. In a cohort study over a period of 12 months Bellelli et al. showed that advanced age, living alone, cognitive impairment, delirium and poor functional status at discharge from a rehabilitation unit were the main predictors of subsequent institutionalization [5]. Multimorbidity and delirium were predictors of rehospitalization. Multimorbidity, often leading to (inappropriate) polypharmacy, is furthermore a predictor for mortality and should be considered at discharge for secondary prevention programs [5]; however, it is unknown whether PIM at admission to geriatric inpatient rehabilitation has an association with rehabilitation outcome. In a recently published study mobility was the main predictor for successful return home after inpatient geriatric rehabilitation [12].

The purpose of this study was to investigate the association of PIM at admission to an inpatient rehabilitation program with the outcome of rehabilitation regarding mobility and quality of life. The hypothesis was that patients taking PIM would have a less favorable rehabilitation outcome regarding mobility and quality of life than patients not taking PIM.

To our knowledge this is the first study to evaluate the effects of PIM at admission to a geriatric inpatient rehabilitation program on rehabilitation outcomes at discharge. In our study population 43% of patients were taking at least one PIM at admission. Patients without PIM had a significantly better quality of life at admission (p < 0.05) and discharge (p < 0.005); however, during rehabilitation the quality of life and mobility of patients with and without PIM increased identically, and the number of PIM was not associated with these rehabilitation outcomes.

In the study of Mann et al. the main group of PIM in Austrian nursing homes was antipsychotic drugs, followed by benzodiazepines, non-steroidal anti-inflammatory drugs (NSAIDs) and opiates [16]. The main groups of PIMs in our study were antipsychotic drugs and benzodiazepines, followed by beta-blockers and diuretics. In the multicenter Austrian study by Mann et al. [16] with 1844 patients, the prevalence of residents taking at least one PIM was 70.3%, whereas in our population only 43.3% of the patients were taking PIM. This difference may arise from the fact that different patient populations were studied. Our patients were community-dwelling persons needing inpatient rehabilitation after an acute hospital stay because of illnesses or an operation, whereas Mann et al. assessed PIM in a nursing home setting. Besides the frequency of PIM, there were no major differences regarding the groups according to the STOPP criteria. In both settings antipsychotic drugs and benzodiazepines were the most prescribed PIM. Fialová et al. showed in an outpatient setting that older people and those living alone were more likely to be taking one or more PIM [7]. Overall, the mean age of the study population was 82.2 years, the majority of patients were women and most lived alone. Polypharmacy was documented in 51% of patients, and 19.5% of patients were taking at least one inappropriate medication. The patients included in our group were slightly younger (mean age 75.5 years) and, in contrast to other studies, more men were taking PIM. This might reflect the fact that men in Switzerland are slightly more prone to a worse health status than women [23]. We found no major differences compared with the results of the study by Gallagher et al. who first published information on the STOPP criteria [9]. The study population of Landi et al. is also comparable to ours, since they included a total of 364 patients [13] and 26% (n = 94) were taking at least 1 inappropriate medication. The most frequently prescribed PIM were benzodiazepines, followed by antihypertensive drugs. Both drug groups (central nervous system drugs and cardiovascular drugs) were also the most prominent PIM in our cohort. Onder et al. showed that during a hospital stay 28.6% of patients were taking inappropriate drugs, 23% took 1 inappropriate drug and 5.6% took 2 or more inappropriate drugs [19]. In this acute hospital setting, the intake of PIM had no significant effect on hospital outcomes, such as mortality, length of stay, or adverse drug reactions.

In our study population, patients who were not taking any PIM had a significantly better quality of life at admission and discharge than patients who were taking PIM. Furthermore, patients who were not taking any PIM were less impaired at admission and discharge and more independent at the end of the rehabilitation. Interestingly, the group with PIM showed a significantly higher comorbidity index at admission than the group who were not taking any PIM, perhaps indicating that a higher illness load may directly lead to an increase in medications and therefore also to taking medications that potentially are not indicated. Nevertheless, it seems that PIM do not have a negative impact on the rehabilitation course, as quality of life and mobility increased identically in both groups from admission to discharge. Patients with PIM generally had lower levels of increased quality of life and mobility than those without PIM. Landi et al. showed that community-dwelling older persons in Italy taking PIM had a significant negative association regarding physical performance [13]. Patients who were taking PIM were less likely to be physically active and showed poorer results in the short physical performance battery. In another study from Runganga et al. in a post-acute outpatient transitional care setting, which is designed to facilitate transitions from hospital to home for older people in Australia [20], polypharmacy (5–9 drugs) was observed in 46.7% of patients. The most prevalent PIM were antidepressants. Polypharmacy was associated with frailty, falls and poor functional outcomes. Finally, regarding function a study from Tosato et al. [24] showed that in geriatric care wards (geriatric evaluation and management units) a decline in physical function was more often identified in patients taking PIM. The decline in function increased with the number of PIM used. Our results are in line with all these findings; our patients with PIM also seemed to be more impaired than patients who were not taking any inappropriate drugs. Although patients with PIM had lower mobility levels at baseline and after rehabilitation, correlation analysis revealed no statistical association between PIM and mobility; however, the mobility of patients taking PIM at hospital entry did not increase to the same outcome level reached by patients who were not taking any PIM despite the fact that they underwent the same rehabilitation program; this might lead to poorer outcomes in the long term. Finally, one could argue that rather than PIM the disease itself may have altered the outcome results, since different diseases may have different disease courses and therefore also different outcomes regarding quality of life and mobility; however, in our study the main diagnoses and diseases were equally distributed, with no statistical differences at baseline. Therefore, we believe that this possible effect of the disease itself would also have appeared in the group of patients not taking PIM.