The Care of the Very Old in the Last Three Days of Life

Abstract

Background:

As life expectancy has increased, the age at which people are dying has also increased. There is limited knowledge of the experience of dying of the very old. We sought to examine the last 3 days of life for the very old, dying in a palliative care unit, focusing upon symptom burden and medications prescribed.

Methods:

A retrospective review of medical records of consecutive patients who died in two inpatient palliative care units. Information collated included demographic and medical information, symptom data, and medications (opioids, benzodiazepines and antipsychotics) administered. Analysis comparing patients aged 80 years and older (cases) and those in the median age range of the treating palliative care units, that is, those aged 50 to 70 years (comparators) were conducted.

Results:

One hundred five cases and 100 comparators were identified. Analysis revealed a significantly shorter length of stay in the cases (13 days) compared to comparators (19 days; p≤0.01).) In the last 3 days, cases received significantly less parenteral morphine equivalents (82.8 versus 170.5 mg, p<0.05), midazolam (12.1 versus 19.1 mg, p<0.05), and lorazepam equivalents (0.9 versus 2.4 mg, p<0.01). Overall, symptom profiles between the groups were similar.

Implications:

The very old appear to have a distinct experience of palliative inpatient care with shorter admissions, and lower requirements for medication. Reasons for lower medication requirements are discussed, and the need for future prospective studies in this area is highlighted. A better understanding of the needs of this population at end of life will enable adequate service planning and improved care.

Introduction

The World Health Organization has reported that many countries are experiencing dramatic increases in population life expectancy.¹ In 2007, around 13 per cent of Australia's population were aged 65 and older, by 2056 this figure is expected to be between 23% and 25%.² In turn, it is projected that the majority of those dying in developed countries such as Australia and the United Kingdom will be aged over 65 years.^3–5 This indicates a growing need for understanding the end-of-life experiences of older people, and in particular, the very old, that is, those aged over 80 years.

While many authors have documented the symptoms of the terminally ill, with pain anorexia, constipation, weakness, and dyspnea being most common,^6–8 few have focused upon the symptoms experienced by the terminally ill older person. Similarly, the pharmacologic symptom management of older dying persons has also received little attention. In studies examining analgesic use in terminal cancer patients, those older patients received lower doses of analgesia^11–13 and were less likely to be prescribed any analgesia at all when compared to a younger cohort.¹⁴

This study seeks to examine the symptom burden, and the common symptom relief medications prescribed in the final 3 days of life of a group of patients aged 80 years, and over, who die in an inpatient palliative care unit. In particular, we sought to test the hypotheses that in an inpatient palliative care unit the very old: (1) have a greater symptom burden than younger counterparts and (2) are prescribed fewer symptom relief medications than younger counterparts.

Method

A retrospective, consecutive review of the medical records of patients who died between September 1, 2008 and August 31, 2009 in two inpatient palliative care units was undertaken. One unit was an 8-bed facility in an acute care hospital and the other a 26-bed unit community-based facility, both located in Melbourne, Australia. Eligible records for study inclusion were those of patients aged 80 years or older (cases) and those aged between 50 to 70 years old (comparators). The comparator age range was chosen as it represented the median age of patients admitted to the two units.

Data collection focused on the last 3 days of life, which will be referred to as Day 1, Day 2 and Day 3, with Day 3 being the day of death. Data collected from the medical records included age, gender, length of stay, pain scores, diagnoses (malignant versus nonmalignant), and comorbidities.

Functional status of patients, prospectively measured using the Australian Karnofsky Performance Status¹⁵ (AKPS) scale, was also recorded. This scale measures patient functional status from 100 (normal, no complaints, and no evidence of disease) through 30 (almost completely bedridden) to 0 (dead).

Symptom presence and severity data were collected from routine prospective recording of symptoms using the Palliative Care Outcome Collaboration (PCOC)¹⁶ data collection undertaken in both units. Symptoms are recorded according to patient self-report or by proxy if the patient is unable, and include difficulty sleeping, appetite problems, nausea, bowel problems, breathing problems, fatigue, and pain. For patients unable to report symptom intensity, formal nonverbal assessment tools are not routinely used. Instead the clinical staff member makes the assessment which is then confirmed (or otherwise) by another staff member. Both staff members are directly involved in the patient care. Symptom severity is rated by patients on a numerical scale from 0 to 10, with 0 indicating absence of the symptom and 10 indicating the symptom is at its most severe. These are data routinely collected as part of clinical care.

The use and dose of opioids, benzodiazepines, and antipsychotics in the final three days of life were documented from medication charts. For comparison purposes, parenteral morphine equivalent doses were calculated for all opioids administered.¹⁷ As it was anticipated that the majority of inpatients would receive midazolam, these doses were noted. For benzodiazepines other than midazolam, an equivalent lorazepam dose was calculated¹⁸ as this was the second most commonly used benzodiazepine within the cohort. There are no agreed upon dose equivalency equations for antipsychotics; hence the doses were reported for individual drugs.

Descriptive statistics including the means, standard deviations, and percentages were used to describe the sample. Independent sample t tests were conducted to compare the case and comparator groups when the tested variable was continuous (age, length of stay, and medication doses). χ² analysis was conducted to compare case and comparator groups when the tested variable was dichotomous (gender, malignant versus nonmalignant diagnosis). A value of p≤0.05 was considered statistically significant. Data analysis was performed using the SPSS Statistics 17.0 (release August 23, 2008; SPSS, Chicago, IL).

Results

One hundred five cases and 100 comparators were identified, with a mean age (standard deviation [SD]) of 85.59 (4.65) and 62.76 (5.76), respectively. Table 1 lists the demographic and clinical information for both groups.

Table 1.

Demographic and Clinical Characteristics of Cases (n=105) and Comparator (n=100)

Demographic information	Cases (aged≥80) (n=105)	Comparators (aged 50–70) (n=100)
Age
Mean (SD)	85.59 (4.65)	62.76 (5.76)
Range	80–100	50–70
Gender, n (%)
Male	50 (47.6)	50 (50)
Female	55 (52.4)	50 (50)
Length of stay^a
Mean (SD)	13 (16.45)	19 (18.33)
Range (days)	0–108	0–73
Diagnosis^a n (%)
Malignant	74 (70.5)	89 (89)
Nonmalignant	31 (29.5)	11 (11)
Number of comorbidities^a
Mean (SD)	1.93 (1.28)	1.14 (0.45)
Range	1–3	1–6
AKPS Mean (SD)
Day 1	23.49 (12.24)	27.67 (15.31)
Day 2	20.60 (10.59)	22.56 (12.12)
Day 3	13.15 (8.87)	11.43 (9.62)

Significant difference between groups (p≤0.001).

AKPS, Australian Karnofsky Performance Status; SD, standard deviation.

The cases had a significantly shorter length of stay (mean=13 days) compared to the comparator group (mean=19 days; p=0.015). Upon comparing malignant versus nonmalignant diagnosis, significantly more cases (n=31) had a nonmalignant diagnosis, compared with the comparator group (n=11; p=0.001). Cases were noted to have significantly higher total number of comorbidities than comparators (p<0.001), including cardiac failure (n=11), renal failure (n=9), respiratory failure/chronic obstructive pulmonary disease (n=6), and hepatic failure (n=2). Of note, cases had diagnoses of dementia (including Alzheimer's disease [n=11]) and delirium [n=3] while no comparators had these disorders.

Overall, the symptom severity scores between the groups were similar. The cases reported significantly more appetite problems on Day 1 and Day 2, more bowel problems on Day 2, and more fatigue on Day 1 and 2 (Table 2). The comparators reported significantly more nausea on Day 2 and 3 when compared to their older counterparts. Pain was not reported to be severe problem in either group and no significant difference was found between the two groups. The average pain scores appeared to be lowest on Day 3 in both groups, which was also the day of death.

Table 2.

Symptom Presence and Average Symptom Severity Scores of Both Groups

	Cases (aged≥80) (n=105)	Comparator (aged 50–70) (n=100)
Symptoms	Mean (SD)	Mean (SD)	p
Difficulty sleeping
Day 1	0.93 (1.94)	0.77 (1.64)	0.572
Day 2	0.81 (1.73)	0.90 (2.21)	0.742
Day 3	0.38 (1.78)	0.49 (1.23)	0.691
Appetite problems
Day 1^a	3.75 (3.74)	2.28 (3.44)	0.009
Day 2^a	4.21 (4.10)	1.88 (3.26)	≤0.001
Day 3	2.65 (3.93)	2.09 (3.79)	0.445
Nausea
Day 1	0.34 (1.41)	0.81 (2.03)	0.082
Day 2^a	0.06 (0.35)	0.57 (1.56)	0.003
Day 3^b	0	0.45 (1.69)	0.046
Bowel problems
Day 1	2.32 (3.24)	1.70 (2.37)	0.173
Day 2^a	3.30 (3.55)	1.61 (2.56)	≤0.001
Day 3	2.02 (3.06)	1.20 (2.66)	0.139
Breathing problems
Day 1	2.71 (3.13)	1.94 (2.69)	0.087
Day 2	3.07 (3.32)	2.38 (3.06)	0.149
Day 3	2.71 (3.21)	2.31 (3.37)	0.517
Fatigue
Day 1^b	5.78 (3.33)	4.75 (3.51)	0.050
Day 2^b	5.32 (3.75)	4.10 (3.82)	0.031
Day 3	3.91 (4.42)	2.44 (4.00)	0.071
Pain
Day 1	2.70 (2.90)	2.00 (2.32)	0.082
Day 2	2.38 (2.57)	2.24 (2.59)	0.726
Day 3	1.50 (1.81)	1.84 (2.99)	0.474

Significant statistical difference with p≤0.01.

Significant statistical difference with p≤0.05.

SD, Standard deviation.

The most commonly prescribed opioid, benzodiazepine, and antipsychotic used were morphine, midazolam, and haloperidol, respectively (Table 3). On average, the cases received lower doses of opioids compared to the comparator group, in the last three days of life. These differences were significant on Days 1 and 2 and in total dose over the 3 days, with the comparators receiving on average 87.74 mg more of parenteral morphine equivalent compared to the older group (Table 3).

Table 3.

Comparison of Average Medication Doses Received between Groups

	Medication and doses
Day	Parenteral morphine equivalent	p	Midazolam	p	Lorazepam equivalent	p	Haloperidol	p
Day 1 (mg/24 hrs)		0.024		0.129		0.067		0.163
Cases	29.87		3.26		0.47		0.62
Comparator	61.88		4.89		0.89		1.05
Day 2 (mg/24 hrs)		0.008		0.011		0.002		0.151
Cases	32.37		4.84		0.31		0.67
Comparator	67.85		8.50		1.08		1.19
Day 3 (mg/24 hrs)		0.161		0.179		0.136		0.607
Cases	20.51		4.04		0.18		0.44
Comparator	40.77		5.74		0.41		0.57
Total over 3 days (mg/72 hrs)		0.027		0.022		0.008		0.112
Cases	82.75		12.14		0.95		1.72
Comparator	170.49		19.13		2.38		2.80

Analysis of prescription of benzodiazepines revealed the average midazolam dose received over the 3 days by the cases was significantly lower (12.14 mg) compared with comparators (19.13 mg, p=0.022).

Following conversion of all other benzodiazepines, except midazolam, to lorazepam equivalents, it was apparent the cases received significantly lower total doses of lorazepam equivalents over the 3 days (0.95 mg), compared to the comparators at the end of life (2.38 mg, p=0.008). The cases received less haloperidol on any given day and also lower total average dose over the last 3 days of life when compared to their younger counterparts; however, this difference did not reach statistical significance (Table 3).

Discussion

This study provides preliminary insights into the care of older patients, their symptom profiles, and symptom management and provides comparisons to the experiences of younger patients, in inpatient palliative care units.

Several significant differences were apparent when comparing the care of an older group with a younger group in their last 3 days of life. The older patients were more likely to have a shorter length of stay, a nonmalignant diagnosis, and multiple comorbidities. While symptom profiles were generally similar between the groups, with only certain symptoms found to be significantly different, the medication prescribed to the older group, including opioids, benzodiazepines, and antipsychotics was significantly less compared to that prescribed to the younger group.

The symptoms and severity reported among patients in the current study are similar to findings reported elsewhere.^7,19–21 Symptom presence and severity in the dying are most likely to be caused by multiple and interacting factors, rather than a single factor.^8,22 These include age,^8,22 gender,⁸ general medical condition,²² performance status,^8,22 tumor location, and medication prescribed.²² The symptom profile in this study between the dying older and younger groups was in general, similar. Only appetite problems, bowel problems, and fatigue were significantly higher in the older population, while nausea affected their younger counterparts more severely. Studies that have considered age-related differences in symptom prevalence found symptoms of pain and dysphagia decreased with age,¹⁰ while urinary incontinence,^9,10,21 fecal incontinence,⁹ and depressed mood^21,23 increased. This relatively similar profile of symptoms between the two groups may be explained by the theory of a “terminal common pathway”^20,24 in which symptoms are thought to converge at the end of life, regardless of age and diagnosis.

Confirming the findings of others, morphine, midazolam and haloperidol were the most commonly prescribed opioid,^25,26 benzodiazepine,^23,27–30 and anti-psychotic^27,28,31,32 in our study, with similar dosage ranges prescribed.^{13,25,28,30,31,33–35} Opioid doses prescribed to the older population were significantly less, the average over 3 days was 87 mg less of parenteral morphine equivalents. Of note, despite this, there were no significant differences in pain severity between the two groups.

Two studies have examined the effect of age upon opioid dosing.^11,13 Although neither study provides information about the pain intensity of the patients, both reported that age was inversely related to opioid dose.^11,13 Bernabei et al.¹⁴ found that patients aged 85 years and older were less likely to receive morphine and other strong opiates. Furthermore, these authors reported that very old patients were less likely to be prescribed any analgesia, even when daily pain was reported.¹⁴

There are a number of possible reasons to account for the differences in opioid prescribing practices in the oldest patients. Some have suggested that the older patients require less opioid to achieve an analgesic effect equal to their younger counterparts.³⁶ Studies on acute postoperative pain management support this theory where the usage of patient-controlled analgesia was inversely related to age, and further conclude that age is the best predictor of morphine dose.^37,38 Changes in pharmacokinetics and pharmacodynamics, for example, reduced renal clearance in older patients, are thought to account for this, at least in part.³⁹ Increased permeability of the blood–brain barrier has been reported in older people and may explain the lower opioid doses.⁴⁰

Older patients may have higher pain thresholds and do not require large doses of opioids. This theory, however, remains controversial with inconsistent findings relating to age and pain levels. Indeed, pain thresholds have been variably reported to decrease, increase, and remain unchanged with increasing age.⁴¹ Another possible explanation for reduced opioid doses in older people may reside in patient related factors, including multiple comorbidities, stoicism,^23,38 and attitudes toward pain and pain medication.^14,22,42 Closs et al.⁴² reported that older patients referred to community palliative care believed that analgesics should only be used when pain is intolerable and that their complaints of pain cause medical staff to perceive them as “bad patients” or make them become impatient. Finally, the reduced opioid doses prescribed in older patients may be explained by physician and medical staff related factors. These may include a lack of communication regarding pain management⁴³ and physicians' possible reluctance in prescribing opioids to the oldest old.⁴⁴ A study conducted by Morrison and Siu⁴⁴ on survival and care received by demented older patients in an acute setting reported only nine out of 38 (24%) such patients with hip fractures received a standing order of analgesics.

Of note, the older group in our study also received less midazolam, lorazepam equivalents, and haloperidol over the last 3 days of life. This has not been previously documented in the literature, although the findings from others studies indicate similar trends. For example, Henderson et al.²⁹ studied the use of benzodiazepines among hospice patients (n=100) in the United Kingdom and reported that benzodiazepine use peaked in the last phase of life, and that younger patients were more likely to be prescribed benzodiazepines. Those prescribed benzodiazepines had a median age of 70 years, compared with 80 years for those not prescribed such medications. Henderson et al.²⁹ attributed the higher doses of benzodiazepines administered to younger patients to higher levels of anxiety experienced by this group, citing external stressors such as young families and work, which are less common for older patients. Similarly, Morita et al.³⁴ reported younger terminally ill patients were more likely to receive high doses of midazolam.

The higher doses of benzodiazepines in the younger patient group found in the current study may be influenced by the higher doses of opioid prescribed. A higher dose of benzodiazepine might be needed to achieve an effect, since cross-tolerance between opioids and benzodiazepines has been described in patients on opioid maintenance therapy.^45,46 The changes in fat distribution and fat to muscle ratios in older people may have implications for the volume of distribution of lipophilic drugs, notably benzodiazepines.^47,48 A further explanation for the differential doses of benzodiazepines may reside in age-related alterations to benzodiazepine receptors in the central nervous system, increasing sensitivity to the drug in those older.⁴⁹ An increased sensitivity can potentially lead to unwanted sedation, unsteadiness, memory loss and disinhibition.⁴⁹ These factors, and the likelihood of older patients already on polypharmacy, might suggest lower doses of benzodiazepines and antipsychotics should be prescribed. This area has received little attention and is deserving of further prospective studies.

In this study, the incidence of delirium in the final 3 days of life was not specifically recorded. It might be expected that delirium would be more common in an older cohort who have increased risk factors for delirium, namely age, comorbidities and preexisting cognitive impairment.⁵⁰ The doses of benzodiazepines and antipsychotics, both frequently used in the management of delirium symptoms, were however, lower in the older group and requires further investigation.

There are several limitations to this study that should be noted. First, the study was conducted in two specialized palliative care units and thus not able to be generalized to the general older population. Second, due to the retrospective nature of this study, missing drug and symptom data were encountered. However, the missing data were minimal including only one record without a drug chart, and four records missing the symptom charts. In addition, there were incomplete data for a number of patients on the day of death, particularly if death occurred in the early morning. Charting of performance status, and symptom presence and severity may become a low priority when nurses are potentially busy managing both the dying patient and distressed family members. A further limitation of this study is that the symptom presence and severity was often recorded by nurses and physicians not by the patient and is therefore subject to observer bias. In the last 3 days of life, we expect most patients were unable to nominate symptom severity themselves, although details of this were not recorded. The increased likelihood of patient cognitive impairment and therefore the staff scoring of symptoms on the last day of life may account for the lower pain scores recorded at that time. It is important to note that while the study was retrospective in design, the recording of functional status, symptoms, and medication use was prospective as part of daily patient care.

Finally, this study focused on opioids, benzodiazepines and anti-psychotics only for symptom management and this is partly due to the ease of comparing the drugs with their equivalent doses. To conclude that age is the sole reason to differences in medication doses prescribed is premature. Other confounding factors, such as tumour primary site, presence or absence of metastasis, use of adjuvant medications, and other comorbid conditions could potentially influence symptoms and opioid doses. The role of these factors deserves investigation in future studies designed specifically to answer these questions. Importantly, the prescription of medication does not describe the full range of nursing, medical, and psychosocial intervention taken in palliative care units. The limitations of medication as a surrogate for a complex form of care are obvious. However, medications were adapted for the purpose of this study as an objective reproducible indicator of one aspect of care delivered. Despite these limitations, this preliminary study establishes that the common symptom relief medications are prescribed to older dying patients in lower doses than younger patients. The reasons for this must be the subject of future studies specifically designed to answer this question.

In conclusion, this preliminary study suggests that there are similarities in the experiences and management of the dying old when compared to a younger population, in particular, the likely symptom profile. There are however, important differences with the oldest patients having a shorter inpatient palliative care stay, higher non-malignant diagnoses, and substantially lower doses of symptom management medications prescribed. Future prospective studies are required to examine in detail the adequacy of symptom relief, medication requirements, and, in particular, examine the factors influencing prescription of such medications in this growing patient cohort.

Footnotes

Author Disclosure Statement

No competing financial interests exist.

References

Davies

. What are the palliative care needs of older people, how they might be met? Copenhagen, 2004. WHO Regional Office for Europe (Health Evidence Network report www.euro.who.int/Document/E83747.pdf. 2009 July 29.

Commonwealth of Australia: Fact Sheet 15—Population Projections, 2010. Department of Immigration and Citizenship, National Communications Branch: Canberra.

Ahmad

, O'Mahony

. Where older people die: A retrospective population-based study. Q J Med, 2005; 98:865–870.

Australian Bureau of Statistics. Deaths AustraliaCat no 3302.02008.

Gomes

, Higginson

. Where people die (1974–2030): Past trends, future projections and implications for care. Palliat Med, 2008; 22:33–41.

Chiu

, Ho

, Chen

. Prevalence and severity of symptoms in terminal cancer patients: A study in Taiwan. Support Care Cancer, 2000; 8:311–313.

Potter

, Hami

, Bryan

, Quigley

. Symptoms in 400 patients referred to palliative care services: Prevalence and patterns. Palliat Med, 2003; 17:310–314.

Walsh

, Donnelly

, Rybicki

. The symptoms of advanced cancer: Relationship to age, gender, and performance status in 1,000 patients. Support Care Cancer, 2000; 8:175–179.

Addington-Hall

, Altmann

, McCarthy

. Variations by age in symptoms and dependency levels experienced by people in the last year of life, as reported by surviving family, friends and officials. Age Ageing, 1998; 27:129–136.

10.

Teunissen

, Wesker

, Kruitwagen

, de Haes

, Voest

, de Graeff

. Symptom prevalence in patients with incurable cancer: A systematic review. J Pain Symptom Manage, 2007; 34:94–104.

11.

Hall

, Gallagher

, Gracely

, Knowlton

, Wescules

. The terminal cancer patient: Effects of age, gender, and primary tumor site on opioid dose. Pain Med, 2003; 4:125–134.

12.

Vigano

, Bruera

, Suarez-Almazor

. Age, pain intensity, and opioid dose in patients with advanced cancer. Cancer, 1998; 83:1244–1250.

13.

Bercovitch

, Waller

, Adunsky

. High dose morphine use in the hospice setting—A database survey of patient characteristics and effect on life expectancy. Cancer, 1999; 86:871–877.

14.

Bernabei

, Gambassi

, Lapane

, Landi

, Gatsonis

, Dunlop

et al. Management of pain in elderly patients with cancer. JAMA, 1998; 279:1877–1882.

15.

Abernethy

, Shelby-James

, Fazekas

, Woods

, Currow

. The Australia-modified Karnofsky Performance Status (AKPS) scale: A revised scale for contemporary palliative care clinical practice [ISRCTN81117481] BMC Palliat Care, 2005; 4:7.

16.

Palliative Care Outcome Collaboration (PCOC). Education handouts: A resource for services. Version 1.8. http://chsd.uow.edu.au/pcoc.html. 2010 April 30.

17.

Twycross

, Wilcock

. Palliative Care Formulary. Nottingham: Palliativedrugs.com Ltd., 2007.

18.

Therapeutic Guidelines: Psychotropic, 6th. Melbourne: Therapeutic Guidelines Limited, 2008.

19.

Klinkenberg

, Willems

, van der Wal

, Deeg

DJH

. Symptom burden in the last week of life. J Pain Symptom Manage, 2004; 27:5–13.

20.

Seah

STA

, Low

, Chan

. Symptoms and care of dying elderly patients in an acute hospital. Singapore Med J, 2005; 46:210–214.

21.

Teunissen

, de Haes

, Voest

, de Graeff

. Does age matter in palliative care? Crit Rev Oncol Hematol, 2006; 60:152–158.

22.

Morita

, Tsunoda

, Inoue

, Chihara

. Contributing factors to physical symptoms in terminally-ill cancer patients. J Pain Symptom Manage, 1999; 18:338–346.

23.

Retornaz

, Seux

, Sourial

, Braud

, Monette

, Bergman

et al. Comparison of the health and functional status between older inpatients with and without cancer admitted to a geriatric/internal medicine unit. J Gerontol Ser A Biol Sci Med Sci, 2007; 62:917–922.

24.

Vainio

, Auvinen

, Ahmedzai

, Benton

, Walsh

, Higginson

et al. Prevalence of symptoms among patients with advanced cancer: An international collaborative study. J Pain Symptom Manage, 1996; 12:3–10.

25.

Bailey

, Ferguson

, Williams

, Woodby

, Redden

, Durham

et al. Palliative care intervention for choice and use of opioids in the last hours of life. J Gerontol Ser A Biol Sci Med Sci, 2008; 63:974–978.

26.

Nauck

, Ostgathe

, Klaschik

, Bausewein

, Fuchs

, Lindena

et al. Drugs in palliative care: Results from a representative survey in Germany. Palliat Med, 2004; 18:100–107.

27.

Drummond

, Peterson

, Galloway

, Keefe

. National survey of drug use in palliative care. Palliat Med, 1996; 10:119–124.

28.

Good

, Ravenscroft

, Cavenagh

. Effects of opioids and sedatives on survival in an Australian inpatient palliative care population. Intern Med J, 2005; 35:512–517.

29.

Henderson

, MacGregor

, Sykes

, Hotopf

. The use of benzodiazepines in palliative care. Palliat Med, 2006; 20:407–412.

30.

Sykes

, Thorns

. Sedative use in the last week of life and the implications for end-of-life decision making. Arch Intern Med, 2003; 163:341–344.

31.

Agar

, Currow

, Plummer

, Chye

, Draper

. Differing management of people with advanced cancer and delirium by four sub-specialties. Palliat Med, 2008; 22:633–640.

32.

Power

, Kearney

. Management of the final 24 hours. Ir Med J, 1992; 85:93–95.

33.

Kohara

, Ueoka

, Takeyama

, Murakami

, Morita

. Sedation for terminally ill patients with cancer with uncontrollable physical distress. J Palliat Med, 2005; 8:20–25.

34.

Morita

, Tei

, Inoue

. Correlation of the dose of midazolam for symptom control with administration periods: The possibility of tolerance. J Pain Symptom Manage, 2003; 25:369–375.

35.

Morita

, Tsunoda

, Inoue

, Chihara

. Effects of high dose opioids and sedatives on survival in terminally ill cancer patients. J Pain Symptom Manage, 2001; 21:282–289.

36.

Mercadante

, Casuccio

, Pumo

, Fulfaro

. Factors influencing the opioid response in advanced cancer patients with pain followed at home: The effects of age and gender. Support Care in Cancer, 2000; 8:123–130.

37.

Macintyre

, Jarvis

. Age is the best predictor of postoperative morphine requirements. Pain, 1996; 64:357–364.

38.

Woodhouse

, Mather

. The influence of age upon opioid analgesic use in the patient-controlled analgesia (PCA) environment. Anaesthesia, 1997; 52:949–955.

39.

Mercadante

, Arcuri

. Pharmacological management of cancer pain in the elderly. Drugs Aging, 2007; 24:761–776.

40.

St-Onge

. Relationship between body composition changes and changes in physical function and metabolic risk factors in aging. Curr Opin Clin Nutr Metab Care, 2005; 8:523–528.

41.

Helme

. Pain in the elderly. Austr J Ageing, 1998; 17:33–335.

42.

Closs

, Chatwin

, Bennett

. Cancer pain management at home (II): Does age influence attitudes towards pain and analgesia? Support Care Cancer, 2009; 17:781–786.

43.

Lovheim

, Sandman

, Kallin

, Karlsson

, Gustafson

. Poor staff awareness of analgesic treatment jeopardises adequate pain control in the care of older people. Age Ageing, 2006; 35:257–261.

44.

Morrison

, Siu

. Survival in end-stage dementia following acute illness. JAMA, 2000; 284:47–52.

45.

Bramness

, Kornor

. Benzodiazepine prescription for patients in opioid maintenance treatment in Norway. Drug Alcohol Depend, 2007; 90:203–209.

46.

Lavie

, Fatseas

, Denis

, Auriacombe

. Benzodiazepine use among opiate-dependent subjects in buprenorphine maintenance treatment: Correlates of use, abuse and dependence. Drug Alcohol Depend, 2009; 99:338–444.

47.

Klotz

. Pharmacokinetics and drug metabolism in the elderly. Drug Metab Rev, 2009; 41:67–76.

48.

Kuk

, Saunders

, Davidson

, Ross

. Age-related changes in total and regional fat distribution. Ageing Res Rev, 2009; 8:339–348.

49.

Bogunovic

, Greenfield

. Use of benzodiazepines among elderly patients. Psychiatr Serv, 2004; 55:233–235.

50.

Innouye

. Delirium in hospitalized older patients. Clin Geriatr Med, 1998; 14:745–764.