Several criteria or indices have been developed to classify appropriate versus inappropriate medications for the elderly including the Beers criteria (or Beers list), the Medication Appropriateness Index, and the Screening Tool of Older Persons' Potentially Inappropriate Prescriptions (STOPP) criteria.^7–9 A limitation of the Beers criteria and these other frameworks is that they are developed specifically for elderly populations, who are only a subset of persons at end of life.² Individuals at the end of life are not all elderly and have varied and dynamic health conditions that are difficult to classify into a simple, yet comprehensive, medication appropriateness index or criterion. A diagnosis of LLE modifies the goals of care towards a more palliative and supportive focus, and subsequently alters the medications that are appropriate or inappropriate.¹⁰

The literature search identified three papers that address the issue of developing specific frameworks or guidelines for identifying appropriate versus inappropriate medication use among persons with LLE.^2,10,11 Of these three articles, two (Holmes et al. and Bain and Weschules) present specific guidelines or frameworks for assessing appropriate medication use at end of life. The remaining paper¹¹ is a letter to the editor that highlights the need to address this issue.

Bain and Weschules¹⁰ examined the feasibility of applying the Beers criteria to medication use among older adults receiving hospice care. They developed a list of potentially inappropriate medications based on the Beers criteria and The Hospice Pharmacia Medication Use Guidelines™ seventh edition. The latter is a set of proprietary pharmacotherapy-based guidelines and algorithms specific to hospice care designed as a clinical tool for the pharmacological management of symptoms common at end of life.¹² Twenty-four members of a multidisciplinary panel reviewed the list and assessed the appropriateness of each medication or medication class for hospice patients aged 65 and older. The panelists considered some medications, e.g. propoxyphene, to be inappropriate for all patients; however, some medications, e.g. lorazepam and haloperidol, which were otherwise considered inappropriate according to the Beers criteria, were judged to be appropriate for hospice care. The authors note the need to develop potentially inappropriate medication criteria that are specific to hospice care.

Holmes and colleagues designed a medications model that is specific to persons with LLE.² This model incorporates individuals' remaining life expectancy, goals of care, and time until benefit of medications.² The Holmes et al. prescribing model is less explicit than the Beers criteria and is meant to assist health-care providers in making decisions regarding medication use for persons at the end of life. The Holmes et al. prescribing model does not provide specific guidelines regarding appropriate versus inappropriate medications, but does provide a framework to direct prescribing to avoid inappropriate or ineffective medication use.

Although there is literature examining potentially inappropriate use of medications among the elderly,^13–17 few studies focus specifically on preventive medications and persons with LLE. Our search retrieved only 11 studies, summarized in Table 1, which focus on the appropriateness of medication use among persons with LLE.^17–27

A considerable number of individuals with LLE continue to receive potentially inappropriate medications, yet results vary by study location and population. Fahlman et al.¹⁷ retrospectively applied the Beers criteria to examine the extent of inappropriate medication use among 4602 community-dwelling individuals in the last year of life in the USA. They identified that 44% of participants received at least one potentially inappropriate medication in their last year of life.¹⁷ In contrast, Nicholson et al.²³ identified that 25% of individuals admitted to a hospice in England were taking futile or unnecessary medications according to the admitting physicians.

In order to assess the changes in potentially inappropriate medication use, Currow et al.¹⁹ examined medication use each month from palliative care referral until death. They identified that the total number of medications taken increased as death approached due to a greater number of symptom-specific medications, while medications for chronic diseases decreased. Moreover, two studies analyzed the impact of a palliative care consultation on medication use. Without distinguishing between appropriate and inappropriate medications, Koh and Koo²² identified that the total number of medications taken prior to and after palliative care assessments in Singapore remained unchanged. Suhrie et al.²⁶ applied the Medication Appropriateness Index⁸ to examine medication use by 89 individuals before and after admission to a palliative care unit in Pittsburg, PA. For the purpose of their study, unnecessary medications were defined as those that lacked indication (e.g. prescribing iron for an individual without an iron deficiency), lacked effectiveness (e.g. prescribing a drug that has greater side effects than an alternative drug), or were duplicate medication. They identified that the use of unnecessary medications declined from a mean of 1.7 to 0.6 medications per person after admission to the palliative care unit.²⁶

To more precisely report on the appropriateness of medication use, selected studies restricted their analysis to a specific condition or specific preventive medication. The literature search retrieved four studies that examined specific conditions: advanced dementia^18,21 and terminal cancer.^20,24 Holmes et al.²¹ examined medication use in 34 persons with advanced dementia who were registered in a palliative care program and assessed appropriateness using a Delphi panel of 13 geriatricians. Twenty-nine percent of persons were receiving a medication that was classified as never appropriate, including HMG-CoA (3-hydroxy-3-methyl-glutaryl coenzyme A) reductase inhibitors (statins), acetylcholinesterase inhibitors, estrogen, and clopidogrel, and 21% were receiving medication that were rarely appropriate.²¹ Blass et al.¹⁸ examined medication use among nursing home residents with advanced dementia (n = 125), in particular how medication use changed as death approached. While the total number of medications prescribed did not vary, the use of opioid analgesics increased, while the use of most other medications decreased, e.g. for cardiovascular disease and dementia.¹⁸ While the study did not determine the appropriateness of the medications, the high rates of medication use are of concern.

Riechelmann et al.²⁴ retrospectively examined 372 persons with advanced/incurable malignancies who were registered with a palliative care program. The study identified that 22% of individuals were receiving at least one unnecessary or duplicate medication. Fifty-six percent of unnecessary medications were statins and 88% of duplicate medications were benzodiazepines.²⁴ In comparison, Fede et al.²⁰ reported that 24% of individuals with advanced cancer in Sao Paulo, Brazil, were receiving unnecessary medication according to the admitting physician.

Two studies (Silveira et al.²⁵ and Tanvetyanon and Choudhury²⁷) focused on statin use in persons with LLE. Silveira et al.²⁵ used a matched case–control study design to examine whether individuals with a recognizable life-limiting condition would have statins discontinued more often and more quickly than a control group. They compared statin discontinuation among persons with a recognizable life-limiting condition, as defined by the Veterans Health Administration Palliative Care Diagnoses Index, to those who died of other causes, matched on age, sex, co-morbid conditions, and socioeconomic status. Silveira et al.²⁵ identified that 51% of cases and 64% of controls received statins until death. There was no statistically significant difference in the timing of discontinuation between cases and controls. Tanvetyanon and Choudhury²⁷ examined patterns of statin discontinuation among 47 persons who were diagnosed with stage IIIb or IV lung cancer and were taking statins at the time of diagnosis. Eighty-nine percent of persons received at least one prescription renewal after diagnosis. Statins were eventually discontinued in 53%. The median time of discontinuation was approximately 250 days after diagnosis.²⁷

The 11 studies examining appropriate versus inappropriate medication use among persons with LLE suggest that a considerable proportion of individuals at the end of life continue to take chronic disease prevention medications and other potentially inappropriate or unnecessary medications. However, there is an indication that as death approaches the shift from curative to palliative goals of care translates into some change in medication use. Beyond these 11 studies that report on the prevalence of unnecessary and inappropriate medication at end of life, two additional papers focus on discontinuing preventive chronic disease medications, and a further six on adding palliative preventive medications.

The literature search identified only two papers with a specific focus on discontinuing chronic disease preventive medications among persons with LLE.^1,3 Individuals in the general population are prescribed long-term medications such as statins to control cholesterol levels, anti-hypertensives to control blood pressure, and bisphosphonates to maintain bone density. However, if a person is diagnosed with a terminal disease, the benefit of these long-term preventive medications may become negligible, and the potential for harm may outweigh the benefit.²⁸ The safety, efficacy, and benefits of statins, anti-hypertensives, and bisphosphonates have been demonstrated for the general population through several randomized control trials;^29–31 however, there is limited evidence of their appropriateness among persons with LLE, in part because persons with LLE are rarely included in clinical trials.³ There are changes in pharmacodynamics and pharmacokinetics caused by terminal diseases, which result in medications reacting differently than expected. This leads to potentially greater risk of adverse effects and possible reduction of benefit.²⁸ Time until benefit – the period of time required for a medication to be of value to a person – of preventive medications must also be considered, as it may extend beyond life expectancy. Therefore, there is a need to evaluate prescribing for persons with LLE and to consider discontinuing medications that are no longer beneficial.² The use of statins is a case in point.

Statins are prescribed for the primary prevention – prior to disease occurrence – and tertiary prevention – minimizing effects of a disease that is present and avoiding re-occurrence – of cardiovascular disease.^29–31 If the goal of care is palliative, long-term prevention of cardiovascular disease is potentially unnecessary, may contribute to polypharmacy, and is an additional cost to the person and/or health system.³ Lastly, persons suffering from advanced disease usually have a natural decrease in cholesterol level through weight loss and studies have shown increased risk of adverse effects from statins among persons with LLE.^32,33

In addition to the increased risk of polypharmacy and thus possible adverse events, preventive medication use by persons with LLE has psychological and financial considerations.¹ If a preventive medication is discontinued in an individual with LLE, they may feel that the health-care providers are giving up hope. In contrast, renewing a person's preventive medication may provide false hope. Therefore, continuation or discontinuation of a preventive medication should be clearly discussed with the individual before being put into practice.¹

In summary, the use of chronic medications that reduce the risk of the occurrence or exacerbation of disease many years in the future should be reconsidered for persons whose life expectancy is limited. However, the introduction of new medications that prevent or reduce the risk of disease- or treatment-related adverse events may be indicated. This shift is illustrated in Fig. 2.

The search strategy identified six papers that specifically address the issue of adding medications to the treatment regimen at end of life to avoid anticipated adverse outcomes.^34–39 For example, it is known that bisphosphonates may prevent bone loss associated with androgen-deprivation therapy for prostate cancer and metastases to the bone from breast cancer.³⁴ Beyond a disease-specific preventive focus, symptom-specific preventive medications can also be beneficial. The use of laxatives to counter the constipating effects of opioid painkillers is a case in point.³⁵

Constipation is one of the most common symptoms experienced by persons with LLE.³⁶ Constipation causes abdominal pain, cramping, and nausea and vomiting, and results in substantial declines in quality of life. The prevalence of constipation among patients receiving palliative care has been reported to be as high as 87%.³⁵ Numerous factors contribute to the high prevalence of constipation among persons with LLE, in particular the use of opioids, dehydration, inactivity, difficulty eating, and specific disease-related effects.^35,36 It is clinically recommended that laxatives be prescribed concurrently with a strong opioid in order to prevent opioid-induced constipation.^35–37 Bouvy et al.³⁸ examined concurrent laxative and strong opioid use among 269 elderly persons throughout the Netherlands. They identified that only 37% of these individuals were prescribed a laxative within 5 days of starting a strong opioid. However, Goodman et al.³⁹ examined laxative use specifically among persons who were admitted to 1 of 11 hospices in the UK and found that 74% received laxatives with strong opioids.