When less is more: exploring inappropriate medication use in older people with cancer
thesis
posted on 2017-03-01, 01:05authored byTurner, Justin Paul
The prevalence of cancer increases with age, with over 60% of cancer diagnoses and 70% of cancer mortality occurring in people aged ≥65 years. As life expectancy increases, the number of people in the United States of America aged ≥65 years diagnosed with cancer is expected to increase by 67% between 2010 and 2030. Predictions indicate that by 2030 up to one in five people aged ≥65 years in the United Kingdom will be diagnosed with cancer during their lifetime. As people age, their risk of developing multimorbidity increases. This may lead to use of multiple medications, termed polypharmacy. Older people have a dynamic health status which can change from being robust to frail, potentially impacting the benefit to harm ratio of their medications. Furthermore, the diagnosis of cancer may alter patients’ treatment goals from a focus on life extension to palliative care. As a result of these changes, medications that were once appropriate may become inappropriate. The time at which cancer is diagnosed may represent an opportunity to review patients’ medications to reduce potential for harm.
Polypharmacy in older people with cancer has been associated with chemotherapy discontinuation, reduced survival, toxicity, and the use of potentially inappropriate medications (PIMs). PIMs, defined as medications where the risk of harm outweighs the potential for benefit, have been associated with harms including increased hospitalisation, morbidity and mortality in older people with cancer. Medications that increase the risk of falls can be classified as potentially inappropriate, and included both psychotropics and medications that increase orthostatic hypotension. HMG CoA Reductase inhibitors (statins) have been deemed potentially inappropriate in people with a limited life expectancy, as the goal of therapy is unlikely to be achieved in this setting. Explicit criteria to determine PIMs do not consider an individual patient's clinical situation, therefore appropriateness of medications such as analgesics may change as patients transition from being robust to frail.
The aim of this thesis was to explore potentially inappropriate medication use in older people newly diagnosed with cancer. Specifically, the thesis focused on: polypharmacy; medications that are deemed potentially inappropriate according to Beers 2012 Criteria; falls risk increasing drugs (FRIDs); and HMG CoA Reductase inhibitors (statins).This thesis also explored the use of analgesics that may be appropriate but may be prescribed differently between people who are robust, pre-frail and frail.
Data were collected from the Royal Adelaide Hospital Geriatric Oncology Multidisciplinary Outpatient Clinic. Patients attending the clinic were aged ≥70 years and newly diagnosed with cancer. Prior to their first appointment, each patient completed a structured questionnaire based on the principles of comprehensive geriatric assessment. The questionnaire gathered information relating to current medication use, comorbidities, falls in previous six months, pain (10 point visual analogue scale [VAS]) and clinical information used to determine frailty (weight loss in previous six months, exhaustion [using two questions from the CES-D scale as adapted by Fried], instrumental activities of daily living [IADLs], Karnofsky Performance Scale [KPS] and physical function [SF-36]). Data were verified by a geriatric oncology nurse at the initial appointment and any data not self-reported were recorded by the nurse.
Medication use was assessed as the point prevalent at the patient’s initial appointment, and included prescription medications, over the counter medications and complementary and alternative medications. PIMs were defined using Beers 2012 Criteria, FRIDs were defined using the definition of the National Board of Health and Welfare in Sweden, and analgesic use was defined as use of paracetamol, non-steroidal anti-inflammatory drugs (NSAIDs) or opiates.
Receiver operating characteristic (ROC) curves were produced to determine sensitivities and specificities of a range of polypharmacy cut-points in relation to a variety of adverse events. The optimal cut-point definition for polypharmacy in older people with cancer was determined by Youden’s index.
Unadjusted and adjusted multivariate binary and multinomial logistic regression were used to calculate odds ratios (OR) and 95% confidence intervals (CIs). Multivariate binary logistic regression was used to investigate the factors associated with polypharmacy (≥5 medications) and factors associated with use of Beers Criteria medications. Additionally, multivariate binary logistic regression was used to investigate the association between statin use and pain in patients aged 70-79 and ≥80 years and the association between analgesic use and pain according to frailty status. Multivariate multinomial logistic regression was used to determine factors associated with use of increasing numbers of FRIDs.
Results from the ROC curve analysis included establishing a range of sensitivities and specificities for each polypharmacy cut-point, with high sensitivities and low specificities observed for polypharmacy defined as ≥2 medications, and conversely, low specificities and high sensitivities when the cut-point ≥10 medications was used. Youden’s index identified the optimal cut-points for exhaustion (≥3), falls (≥2), physical function (≥20), KPS (<70) and frailty were 3.5, 5.5, 6.5, 6.5, 6.5 medications respectively.
The prevalence of polypharmacy (≥5 medications) was 57%. In the adjusted analysis polypharmacy was associated with poor physical function (OR 1.13, 95%CI 1.06-1.20) compared to normal physical function, and being pre-frail (OR 2.35, 95%CI 1.43-3.86) and frail (OR 4.48, 95%CI 1.90-10.54) compared to being robust. The prevalence of inappropriate medications defined by Beers Criteria was 27%, and use of a Beers Criteria medication was associated with polypharmacy (≥5 medications) (OR 4.10, 95%CI 2.26-7.44) and being frail (OR 3.05, 95%CI 1.18-7.87) compared to being robust. FRIDs were used by 76% of the cohort, with 31.2% using psychotropics and 76.1% using medications that cause orthostatic hypotension. Use of psychotropics was associated with falls risk factors including history of falls (OR 13.50, 95%CI 2.64-68.94) and frailty (OR 27.78, 95%CI 6.06-127.42). No association was identified between use of medications that cause orthostatic hypotension and falls risk factors.
The prevalence of analgesic use varied according to frailty status, with frail patients having a lower prevalence of NSAIDs, and higher prevalence of opioids than robust patients. When analysing analgesic use by frailty, there was weak evidence to suggest that robust patients had higher odds of analgesic use per unit increase of pain score (OR 1.30, 95%CI 0.995-1.71). This relationship was not significant for frail patients per unit increase in pain score (OR1.08, 95%CI 0.934-1.26). Pre-frail patients displayed a concave relationship, with both low and high pain scores being associated with lower odds of receiving analgesics.
The ROC curve analysis demonstrated a narrow range of optimal polypharmacy cut-points to identify a variety of adverse events in older people with cancer. This result is consistent with previous work in general older people and supports the use of five or more medications as the cut-point to define polypharmacy. This thesis was the first body of work to demonstrate an association between both polypharmacy and PIM use and frailty in older people newly diagnosed with cancer. This is important as frail patients may be at greater susceptibility to a range of adverse events including falls, disability or hospitalisation. Furthermore, frailty can influence treatment decisions in older people with cancer.
The use of FRIDs was highly prevalent. Nearly one third of patients used psychotropic medications, which was associated with risk factors for falling, including frailty and history of falls. Additionally, use of statins was highly prevalent in patients aged ≥80 years, with two in five patients using statins. Despite uncertainty relating to benefits of statin use in patients aged ≥80 years, this thesis demonstrates statin use was associated with higher odds of experiencing moderate to severe pain (≥5 in a 10 point VAS) in patients aged ≥80 years. These results suggest that prescribers are not aware or may underestimate the harms associated with FRIDs or statins in older people with cancer.
The analysis of analgesic use suggested judicious use in robust or frail patients, despite some analgesics being classified as inappropriate by Beers criteria. However the non-linear relationship between pain scores and analgesic use in pre-frail patients warrants further research to ensure appropriate use of analgesics in this vulnerable group.
This thesis generates a solid foundation for future research. The research presented in this thesis was cross-sectional and, therefore, could not determine causality. Further prospective longitudinal research is required to determine causality between PIM use and adverse outcomes. Additionally, further research is required to determine whether deprescribing of PIMs in older people newly diagnosed with cancer can provide clinically important outcomes such as reduction in frailty, falls and pain, or improvements in physical function or quality of life.
In conclusion, this thesis provides important data to raise the awareness of PIM use in older people newly diagnosed with cancer. Evidence is provided to support a clear definition of polypharmacy, and to demonstrate polypharmacy was associated with frailty and multiple comorbidities. A high prevalence of PIM use was observed using a range of metrics, and was associated with higher odds of frailty, pain, falls risk factors, impaired physical function and a higher number of comorbidities. (...)