Introduction

Palliative care for older (or younger) people with dementia should be a straightforward topic. Dementia is an acquired, chronic, usually progressive and incurable disorder, which normally ends in death. So palliative care seems entirely appropriate. There are disputes about the exact prevalence of the different types of dementia, but it seems reasonable to suggest that Alzheimer’s disease is the most common, with a prevalence of 50% to 60%. About 10% to 20% of dementias are vascular in origin; 10% to 20% are dementia with Lewy bodies [1]. The rest are a huge mixture of pathologies, the largest group of which is formed by the different types of frontotemporal degeneration. There is no dispute that the problem of dementia is growing as the population ages. In 2007, the estimates for the United Kingdom (UK) were that about 700,000 people had dementia [2], but this has already risen. In 2010, the global prevalence was estimated to be 35.6 million, but the estimate is that there will be 115 million people with dementia by 2050 [3]. Furthermore, in several studies comparisons have been made with cancer care. It has been demonstrated that the burden of symptoms, even if different in detail, is similar overall between cancer and dementia: both for the person living with the condition and for their informal carers [4, 5]. There are grounds for saying, therefore, that the need for palliative care in dementia is imperative [6, 7].

In which case, why should we even raise a doubt about the idea of palliative care for people with dementia? We shall highlight two reasons. First, the diagnosis of dementia is becoming contentious [8]. Uncertainty about particular types of dementia, even Alzheimer’s disease, is also evident [9]. Secondly, the nature of palliative care is changing, which leads to conceptual dispute over exactly what it entails [10].

In what ways, however, do such academic debates affect the need for palliative care for those who suffer from these conditions? The debates about dementia are pointing us in the direction of a much greater biomedical concentration on the individual. The clarion call of those who take a psychosocial approach to dementia has been, for many years, that we require person-centered care [11]. But this has tended to entail a focus away from biomedicine. If, instead of Alzheimer’s disease, our focus is on the individual brain and the mixture of ways in which its ageing might be manifest and, therefore, the mixture of ways in which we should try to ameliorate the consequent problems, then the biomedical focus is as much on the individual as is the psychosocial. Individual people with dementia require the sort of breadth of attention that characterizes palliative care.

But there is more to be said, because some of this attention will be with curative intent. Whereas the current drug treatments for Alzheimer’s disease (the cholinesterase inhibitors and memantine) are only symptomatic, the next generation of drugs, which are in development, are intended to be disease-modifying [12]. In any case, we already have a number of examples of the (albeit rare) dementias that, in principle, can be cured or at least treated to very good effect: e.g., normal pressure hydrocephalus, cerebral vasculitis, Hashimoto’s encephalopathy, HIV-associated dementia, and dementias caused by other infections, such as syphilis and Lyme disease. So the dichotomy between cure and care, which has been conceptually built into the notion of palliative care, may not seem so relevant in an environment where care is provided by treatments given with curative intent. Hence, it might be that a new model is required, one that is broad enough to capture the new possibilities of treatment and management. A major contender would be the model of supportive care: “… a complete mixture of biomedical dementia care, with good quality, person-centered, psychosocial, and spiritual care under the broad umbrella of holistic palliative care throughout the course of the person’s experience of dementia” [13].

In the remainder of this review, we shall focus on some specific issues in connection with palliative care for people with dementia, before ending with the broader theme, which concerns the overall approach to people with dementia.

Reviews of the Literature

In recent years there has been increasing interest in the theme of palliative care for people with dementia. A review of the literature in 2007 reflected on the lack of good-quality evidence in support of what seems like a good idea; nonetheless, the review outlined an emerging field of expertise [14]. Two reviews, which appeared in 2010, cover much of the literature since the earlier review, including literature that appeared late in 2009 [15, 16]. So, with one or two exceptions, we shall focus the rest of this paper on publications from 2010 or later.

Where are We Now?

The literature shows a number of areas which still need to be investigated in order to improve our understanding of and ability to effect palliative care in dementia. Practitioners and experts have identified four broad areas of importance to take things forward: advance care planning, psychological support (including in bereavement), management of acute events, and terminal care [17]. We shall return to discuss advance care planning, but note that an immediate problem concerns prognostication.

In a population-based study, estimated median survival time from onset of dementia to death was 4.1 years. Those aged 65 to 69 years survived 10.7 years; those aged over 90 years 3.8 years [18]. More recently, in a study of nursing homes around Boston, MA, it was shown that pneumonia, fever, and eating problems were associated with high 6-month mortality rates [19•]. The same study confirmed the lack of good-quality palliative care in such settings, where distressing symptoms and burdensome interventions also seemed more common in people with dementia approaching death, unless they had well-informed relatives or proxies. What we also know is that better palliative care is required in other settings too for people with dementia [20].

In the United States, people with dementia can receive hospice care if they are within the last 6 months of life. But this is hard to predict. In a recent study, however, it was found that almost all hospices and the majority of non-hospice palliative care programs in the US provide care for people with dementia [21]. But people would probably benefit from palliative support earlier in the disease. They would also benefit from more information. A study in the UK showed that people with dementia and their carers do not understand the condition or its likely progression and, furthermore, this lack of information seems to influence decision-making, which can often lead to inappropriate care [22].

Dying in Different Places

People with dementia die in a variety of settings: home, nursing and residential homes, and in hospital. Relatively few die in hospices. In a study of five European countries, it was found that most died in long-term care homes: between 50% in Wales and 92% in the Netherlands [23]. Contrariwise, only 3% of those in the Netherlands died in hospital, whereas in Wales the figure was 46%. Home deaths rarely occurred (3%–5%), but were more common in Belgium (11%).

In a review of the literature concerning end of life for people with dementia living in the community, out of 68 papers selected, only 17% (12) specifically considered dying with dementia at home [24]. The authors commented: “The few studies that did develop dementia-specific tools to guide end-of-life care and outcome measures specific to improve comfort and communication demonstrated both what could be achieved and how much more needs to be done.”

Meanwhile, in a literature review that looked at the accounts of family carers concerning their experience of care homes in connection with the end of life and death of the person with dementia, there was shown to be a high level of unmet need [25]. The specific decisions that are required toward the end of life are difficult and stressful. This is especially so given the emotional strains, for example the experience of anticipatory grief, with which family carers are burdened, often with little support. The authors state that the curative-palliative care split is not helpful under these circumstances. They argue for “comprehensive care,” which would be very similar to the aspirations of supportive care mentioned above.

Of course, this highlights the extent to which palliative care has to be considered over a longer period of time in dementia. In long-term care homes, therefore, palliative care has to be provided by the staff. Specialist palliative care would only be required in specific circumstances. Dealing with behavior that challenges is a routine part of life for staff looking after people with advanced dementia and can be considered an aspect of palliative care. Education of staff to help them deal with this behavior, therefore, is one way in which palliative care might be improved for people with dementia. In a study in Norway, a 2-day staff education program, followed up by monthly support, improved the quality of care and decreased the severity of agitation [26]. Part of the focus of the education was to highlight the importance of person-centered care over against a more biomedical approach to dementia.

Similar concerns emerge in studies that consider the treatment or management of people with dementia who are admitted to acute hospital settings. In a longitudinal cohort study in a London hospital, it was demonstrated that in 617 people aged over 70, and mostly admitted with infections, 42% had dementia [27•]. The prevalence of dementia rose with age. The researchers found that the risk of death increased significantly with the level of cognitive impairment: about 18% of those with mild to moderate dementia had died within 14 days, but 24% of those with moderate to severe dementia had died. They also posed a question as to whether the admissions to the acute hospital were justified given that there was such high short-term mortality. Meanwhile, in a retrospective case note study in a hospital in Dublin, it was shown that people with dementia received different end-of-life care compared with those without dementia [28]. Specifically, people with dementia were less likely to be referred for palliative care interventions, less likely to be prescribed palliative drugs, and their carers were less likely to be involved in decision-making.

Using Antibiotics

For a long while there has been controversy over the use of antibiotics in people with severe dementia. More work needs to be done, because consensus still seems a little way off. Some years ago, a study showed no difference in terms of survival between people with dementia receiving antibiotics and those receiving palliative management only [29]. More recently, it has been demonstrated that patients with dementia and pneumonia, in whom antibiotics were withheld, showed more discomfort than those who had received antibiotics [30]. A study by the same group in the Netherlands, involving 706 patients, suggested that most pneumonias in people with dementia are treated with antibiotics: 69% with the intention of effecting a cure and 8% with palliative intent [31]. Those where antibiotics were withheld (23%) tended to be patients with more severe dementia and pneumonia, and with poorer food and fluid intake, often with dehydration. In other words, physicians seemed to be making nuanced decisions about the person’s overall level of frailty. A more recent paper, again by the same team, has compared discomfort in people dying from pneumonia and in people dying because of problems with eating and drinking (but without pneumonia) [32•]. They demonstrated that dying from pneumonia caused more discomfort than dying from dehydration and that antibiotics lessened discomfort even quite close to death. They acknowledged, however, that these findings need to be replicated in a prospective study, since their analysis was based on a secondary analysis of two other, older prospective studies.

A possible consensus, therefore, might have been that antibiotics are useful as a palliative tool in advanced dementia for the treatment of pneumonia. However, it would have to be recognized that other measures, too, could contribute toward palliation of the discomfort of pneumonia, e.g., antipyretics and analgesia, plus tepid sponging. But this consensus is premature. A prospective study from the US has now shown that antibiotics prolong life but increase discomfort [33•]. The paper suggests, therefore, that clinicians and families have to weigh up whether they are aiming at survival or comfort. The authors also looked at the different routes of administration for antibiotics. It is suggested that oral or intramuscular antibiotics might be indicated if survival is the aim, but they do not see much advantage to intravenous antibiotics, especially if this involves hospital admission. If the person were very near to death, where perhaps comfort is the prerequisite, then other palliative measures would seem more appropriate. Thus, more research is required. But it is worth commenting that the difficulties in this type of research are huge, because of the confounding variables and because of the need for great ethical sensitivity.

Other Medications

We have already mentioned behavior that challenges, which is common as dementia advances and is discussed under the umbrella of behavioral and psychological symptoms of dementia (BPSD). These can be clustered into psychosis, agitation, and mood disorders and are said to occur in up to 80% of people with dementia in long-term care. Guidelines suggest that the management of choice should be to use nonpharmacological means first [34]. Nonetheless, a great variety of medications has been used and some of the best evidence suggests that, particularly when agitation is severe or potentially dangerous, antipsychotics can be useful [35]. Although, however, there is good evidence that these medications can be effective, there is also significant evidence that they do harm. In the UK, such was the concern, the Department of Health commissioned an independent report [36]. The findings suggest that if 1000 people with BPSD were treated with an antipsychotic for 3 months, there would be 91 to 200 patients with behavioral disturbance or 72 patients with psychosis who would show significant clinical improvement. But in addition, according to the estimates of the report, there would be an additional ten deaths, 18 cerebrovascular adverse events, of which about nine would be severe, plus an additional 58 to 94 patients with gait disturbances, although no additional falls or fractures. In the UK, therefore, there is a push by the government to reduce the use of antipsychotics.

Not everyone, however, accepts the official stance against the use of antipsychotics. An alternative view is that, since we know these drugs can work, then (despite the risks) they should be used carefully, especially if our aim is to palliate: “Severe distress should be appropriately palliated using the best and most effective medicines with the overriding goal of reducing suffering. … If old-age psychiatry services move more clearly towards a palliative model as dementia progresses, they may become more able to justify the use of antipsychotic drugs for the express purpose of reducing suffering” [37].

It is certainly true, as we argued above, that BPSD would form a suitable target for a palliative approach in dementia. But here, again, we can see the force of the notion of supportive care. For it may be the case, after all, that some features of BPSD would be susceptible to curative treatment, perhaps using a variety of psychosocial and pharmacological approaches. In addition, complementary therapies might be useful, since there has been some success with aromatherapy, albeit more evidence is required [38].

On the whole, especially at the end of life, a reduction in the number of medications that a person requires is seen as good palliative care. There is, however, limited and sometimes contradictory evidence to guide discontinuation of psychotropic drugs and medications prescribed for the treatment of physical health problems in patients with advanced dementia who are nearing the end of life [39•]. Research is needed to provide guidance on the discontinuation of drug treatment for background conditions. Part of the problem is to do with estimating life expectancy, for this would guide decisions about the use of drugs that are intended to decrease the chances of an event at some stage in the future. For instance, drugs used to protect against cardiovascular or cerebrovascular events, such as aspirin, antihypertensives, or HMG-CoA reductase inhibitors (i.e., statins), may have little usefulness if the person is going to die from dementia in the next few months. But given the lack of certainty in prognosis, it becomes a difficult ethical question as to whether drugs should be discontinued.

What About Pain? Or is it Distress?

It has become commonplace to comment that pain might not be recognized in people with advanced dementia, who cannot communicate their pain in the normal ways. The suggestion is then that, as a consequence, they receive inadequate analgesia. Hence, there has been a large increase in the number of observational assessment tools aimed at picking up pain in people with dementia. Unfortunately, it is not yet possible on the basis of their psychometric properties to recommend any particular instrument [40]. Much of the literature about pain in dementia has recently been reviewed [41]. A question has been raised concerning whether people with dementia experience pain in the normal way. It may well be that, in certain types of dementia, central processing of pain is affected. But it should not be assumed that people with dementia experience no pain. The evidence is that they do experience pain and they respond to treatment.

Another issue concerns the need to distinguish pain from more general distress. Distress can be caused by pain, but is also caused by fear, by hallucinations, by depression, and so on. Hence, given that there are no signs that are pathognomonic of pain, clinicians have to be careful in case they are picking up signs of psychosocial distress rather than pain. To give analgesia for anxiety caused by noise would be inept and inexcusable, if it could be prevented.

Unfortunately, the use of observation instruments does not lessen the likelihood of such a mistake occurring, because they too pick up non-pain causes of distress. Thus, in a recent study of one particular observational pain scale (PAINAD), while it was good at picking up pain (having a sensitivity of 92%) and measuring change when the pain was treated, it also tended to pick up psychosocial distress, having a specificity for pain of only 61% [42]. The conclusion must be that the assessment of distress and pain involves a good understanding of the person with dementia and a fair amount of clinical acumen. Such assessments need to be as broad as possible and should pay particular attention to any previous history of conditions, such as arthritis, that might still be causing pain under certain circumstances, such as when the person is moved.

Artificial Feeding

Over recent years, despite the persisting tendency to use tube feeding in many care homes in the US, there has been a consensus amongst those writing in the field of palliative care for people with dementia that tube feeding is not appropriate. This is largely based on a review from some years ago [43], which was supported by an ethical analysis of the issues [44]. Subsequent guidelines in the UK have accepted this evidence [45]. There has, however, been some dissent [46]. Indeed, it seems very reasonable to point out that dysphagia is not always a terminal event in dementia. And it seems very reasonable to argue that individuals should be assessed carefully and not subject to a blanket ban on using artificial means to provide nutrition and hydration. It also seems reasonable to argue that there should be a greater emphasis on nutrition at an earlier stage in dementia to prevent malnutrition later. But the evidence remains, as shown by a recent systematic review, that there is no evidence of increased survival in patients receiving enteral tube feeding [47••]. The same review found no evidence of benefit in terms of nutritional status or the presence of pressure ulcers (where it is surmised that better nutrition should prevent ulcers). Hence, the review concluded that, despite the large number of people receiving this treatment (e.g., in the US), “there is insufficient evidence to suggest that enteral feeding is beneficial in patients with advanced dementia.” There is also a lack of data concerning the adverse effects of artificial feeding.

With these comments in mind, it is interesting to note the findings of a recent paper, which considered the differences between two nursing homes in the US with a high and a low prevalence of tube feeding [48]. The researchers found that nursing home culture influenced the use of feeding tubes. Homes with a low rate of tube-feeding had several key features, including a physical environment that promoted the enjoyment of food, administrative support, empowerment of staff to value hand-feeding, and shared decision-making processes involving family members. Alternatively, where tube feeding was more common, the researchers commented on the institutional environment, where there were poorly staffed mealtimes and worries about aspiration.

Advance Care Planning

Given that one of the main difficulties in dementia is that people lose their decision-making capacities, it would seem natural to emphasize the possible benefits of advance care planning (ACP). In the UK, ACP has become part of government policy [49]. There is a concern, however, that there is a dearth of good-quality research to demonstrate the benefits of ACP, in particular in dementia where there are specific difficulties [50]. These often center on the timing and content of ACP for people with dementia.

Indicative research is starting to emerge, with two important studies coming from Australia. In the first, when advance care planning was instituted in nursing homes, emergency calls to ambulances and hospital admissions decreased significantly compared with control homes without the intervention [51]. Only a small proportion of the participants in this study had dementia. In the second study, people over the age of 80 years were randomized to receive usual care or facilitated ACP. People who lacked competence were excluded from the study, but nonetheless it showed significantly positive effects in terms of end-of-life wishes being known and followed; it also showed that families of patients who had died (who were in the intervention group) demonstrated less stress, anxiety, and depression [52•]. Questions about when and how to perform ACP remain difficult in dementia. It is generally accepted that it is a process, not a one-off event. It also requires the right sorts of skill and discernment.

Conclusions: An Ethical Framework

In fact, there are a number of features of palliative care that depend on the right approach. This is not only because the subject matter is sensitive, but also because there are worries about the nature of the decisions that have to be made. Most of these decisions, if not all, have an ethical component. At the end of 2009, the Nuffield Council on Bioethics produced a report looking at ethical issues in dementia [53••]. The report sets out an ethical framework to approach issues that arise in the context of dementia care. The framework, with its emphasis on the person and the need for solidarity, to provide support both to the person with dementia and to the family carers, was in favor of the sorts of approach outlined in this paper. Thus, “It is clearly imperative that key elements of the palliative care approach—in particular the focus on caring for the ‘whole person,’ the importance of advance care planning …, the provision of adequate pain relief, the avoidance of inappropriate treatment, and support for family—should be made available in an appropriate way to people who are dying with dementia.”

We end, therefore, as we began, with a call for the right approach to people with dementia. It is too easy to get it wrong, as shown by the stigma associated with dementia and highlighted in the Nuffield Council’s report. The antidote to this is to approach the person broadly. The palliative care approach, which embraces biological, psychological, social, and spiritual aspects of our humanity, is undoubtedly on the right lines. Perhaps the broad and longitudinal view of supportive care (which includes palliative care), from the earliest signs or symptoms through to death and bereavement, yet without the dichotomy of cure or care, would provide the right sort of holism. It is certainly a perspective that would do no harm.