Palliative Sedation in End-of-Life Care and Survival:
A Systematic Review
- Marco Maltoni⇑,
- Emanuela Scarpi,
- Marta Rosati,
- Stefania Derni,
- Laura Fabbri,
- Francesca Martini,
- Dino Amadori and
- Oriana Nanni
+ Author Affiliations
- Corresponding author: Marco Maltoni, MD, I.R.S.T., Via Piero Maroncelli 40, 47014 Meldola (FC), Italy; e-mail: ma.maltoni@ausl.fo.it.
Abstract
Purpose Palliative sedation is a clinical procedure aimed at relieving refractory symptoms in patients with advanced cancer. It has
been suggested that sedative drugs may shorten life, but few studies exist comparing the survival of sedated and nonsedated patients. We present a systematic review of literature on the clinical practice of palliative sedation to assess the effect, if any, on survival.
Methods A systematic review of
literature published between January 1980 and December 2010 was
performed using MEDLINE and EMBASE databases. Search terms
included palliative sedation, terminal sedation,
refractory symptoms, cancer, neoplasm, palliative care, terminally ill,
end-of-life care, and survival. A manual search of the bibliographies of electronically identified articles was also performed.
Results Eleven
published articles were identified describing 1,807 consecutive patients
in 10 retrospective or prospective nonrandomized
studies, 621 (34.4%) of whom were sedated. One case-control study was excluded from prevalence analysis. The most frequent reason for sedation was
delirium in the terminal stages of illness (median, 57.1%; range, 13.8% to 91.3%). Benzodiazepines were the most common drug category prescribed. Comparing
survival of sedated and nonsedated patients, the sedation approach was not shown to be associated with worse survival.
Conclusion Even if
there is no direct evidence from randomized clinical trials, palliative
sedation, when appropriately indicated and
correctly used to relieve unbearable suffering,
does not seem to have any detrimental effect on survival of patients with terminal cancer. In this setting, palliative sedation is a medical intervention that must be considered as
part of a continuum of palliative care.
INTRODUCTION
There is a widespread consensus that physicians have an ethical obligation to relieve pain and other distressing symptoms
in patients with advanced cancer.1 Despite the progress made in palliative medicine in terms of
symptom control, there are still many patients who have intractable
symptoms, because the treatment is either ineffective
or intolerable. In these circumstances, sedative
drugs are commonly prescribed to control refractory symptoms and relieve
unbearable suffering in those with advanced cancer.
However, palliative sedation (PS) at the end of life has aroused concern in the same way as that of opioids.2 Some authors have suggested that these drugs may shorten life, thereby confusing the boundaries between PS and euthanasia.3 Indeed, PS has been dubbed by some as slow euthanasia or terminal sedation, both terms suggesting that patients' lives are
shortened by treatment.3 This has sparked a wide debate in the palliative care world.
Here we present the results of a systematic review of literature published over the past 30 years concerning the clinical practice of PS. The aim of this review was to evaluate the effect of
sedation on survival, when appropriately indicated and correctly used
to relieve unbearable suffering. In particular, we
wanted to determine if there is a significant
difference in survival between sedated and nonsedated patients and if
the use
of sedatives is associated with anticipation of death.
METHODS
Study Design and Search Strategy
According to the review protocol approved by the medical scientific committee of our institute (Istituto Scientifico Romagnolo per lo Studio e la Cura dei Tumori), a systematic review of literature published between January 1980 and December 2010 was performed using MEDLINE and EMBASE databases. Because the
global definition of PS at the end of life is not generally accepted, we adopted the following search terms: palliative sedation, terminal sedation, refractory symptoms,
cancer, neoplasm, palliative care, terminally ill, end-of-life care, and survival. A manual search of the bibliographies of electronically identified articles was also performed. Studies were included in the analysis if they reported on the length
of survival of terminally ill adult patients with cancer referred or not referred for PS therapy. The previous end point was crucial in excluding or including the articles in our systematic literature review (references not included in
the review are listed in the Appendix,
online only). Only articles published in English were selected. Case
studies, letters, reviews, editorials, and studies
focusing on euthanasia and assisted suicide,
ethical aspects, or opinions were excluded, as were articles that did
not report
the length of survival of each sedated and nonsedated group of patients. The following study characteristics were recorded: first author, year of publication, sample size, type of study (randomized clinical trial or prospective, retrospective, or cohort study), study location (hospice, hospital, home
care), number of patients sedated, reasons for sedation, length of PS (days), type of sedative used, mode of sedation (primary/secondary, intermittent/continuous, proportional/sudden, mild/deep), mean and/or median length of sedative use (days), and mean and/or median overall survival (days).
Selection of Trials and Data Collection
Two reviewers (M.M., E.S.) independently assessed the eligibility of the studies identified by the search. The same reviewers extracted the data independently
using a data collection form predefined in the study protocol. All data
were checked for internal consistency, and any
disagreements in interpretations were resolved
by a discussion and consensus approach. All selected articles had to
present
a reliable measurement of outcome. Length of survival of sedated and nonsedated patients was collected and tabulated for each case series. The methodologic quality of each study was assessed according to the criteria proposed by Hawker et al.4 Each part of
the study was appraised as good, fair, poor, or very poor. PRISMA
(Preferred Reporting Items for Systematic Reviews and Meta-Analyses)
guidelines for building reviews were strictly
followed.
Statistical Methods
The main outcome measure for this systematic review was length of
survival for sedated and nonsedated patients. Because the study
designs, participants, interventions, and reported outcome
measures varied markedly, we decided to focus on
describing the studies, their results, and their limitations and on a
qualitative
synthesis rather than combining the data in a
meta-analytic statistical approach.
RESULTS
Fifty-nine articles were identified through an electronic and manual literature search strategy summarized in Figure 1. Of
these, 48 were excluded for the following reasons: reviews (n = 4),
guidelines (n = 2), case reports (n = 2), surveys (n
= 8), letters (n = 1), ethics articles (n = 3), and
articles not presenting survival comparisons between sedated and
nonsedated
patients (n = 28; Fig 1; excluded articles are described in the Appendix, online only).
Fig 1.
Of the 11 articles that matched the inclusion criteria,5–15 there were no randomized trials; seven studies were retrospective,6,7,9–13 and four were prospective.5,8,14,15
All involved consecutive patients except for one prospective matched
cohort study presented by our group, in which the sedated
group was consecutively recruited, and the
nonsedated group was simultaneously recruited and matched for age, sex,
reason
for hospice admission, and Karnofsky performance status.15 Although most of the studies were designed to evaluate the prevalence and type of symptoms in terminally ill patients, the study methodologies were found to be heterogeneous in terms of inclusion criteria, data collection, and care setting. Sample size varied considerably between studies, with 76 patients
in the study by Fainsinger et al7 and 548 in that by Muller-Busch et al.9 Ten studies involved hospitalized patients from palliative and/or acute care units, and one involved patients receiving home
care.5 Study characteristics are listed in Table 1.
View this table:
Table 1.
Study Characteristics
There was great variability in the number of patients receiving sedation (range, 14.6% to 66.7%). Of
1,807 consecutive patients enrolled onto 10 studies, 621 (34.4%) were
sedated. Although delirium was the most common indication
for PS, there was wide interstudy variability
(range, 13.8% to 91.3%). Other frequent reasons for sedation were
dyspnea (range,
8.7% to 63.0%) and pain (range, 9.5% to 49.2%). Of the 10 studies reporting the main refractory symptoms requiring sedation (Fig 2),5–9,11–15 seven reported psychological distress as one of the main reasons for prescribing PS.6,9,11–15 Mean or median duration of sedation varied from 0.8 to 12.6 days. We also found a significant difference in type of
drug used, not only among countries but also among care units in the
same country. Midazolam was the most common drug prescribed
in nine of the studies (Fig 3).6–8,10–15
Psychotropic drugs were also frequently used, sometimes in conjunction
with benzodiazepines. However, they were the most
favored drug category in only two reports, one
citing haloperidol, and the other, chlorpromazine and lorazepam
(administered
intravenously and/or subcutaneously).
PS characteristics are listed in Table 2. Proportional sedation was the most common method of drug administration; few patients received sudden sedation (deliberately rapid loss of consciousness, inducing deep sleep). Furthermore, only four studies reported results of sedation in terms of relief of distress.7,8,12,14 In most studies, survival was defined as the number of days from hospice/hospital admission or from the start of home care to death. Survival from the start of sedation therapy was also reported in two studies.6,8 Median survival of sedated and nonsedated patients varied from 7 to 36.5 days and from 4 to 39.5 days, respectively; this was not statistically
different between the two patient groups (Table 3).
View this table:
Table 2.
Palliative Sedation Characteristics
View this table:
Table 3.
Survival in Days From Admission: Sedated Versus Nonsedated Patients
DISCUSSION
Despite the huge progress made in palliative medicine in terms of
symptom control, many are intractable (refractory symptoms), either
because the treatment is ineffective or because the treatment
itself is intolerable.1,2 PS, aimed at offering relief from unbearable suffering, is therefore the only reasonable option left to control these symptoms. A recent
systematic review addressed the feasibility of PS in the residential setting, showing it to be a realistic treatment option for those who choose to die at home.16 In a survey of European oncologists, Cherny et al17 reported that although PS should be an integral part of the professional skills of medical oncologists, few of those interviewed felt sufficiently confident in their ability to manage PS. Some authors have suggested a negative impact
of PS on survival, becoming a kind of slow euthanasia.3,18 However, the European Association for Palliative Care Ethics Task Force has clearly stated that PS is a medical intervention,
totally different from euthanasia in aim, procedure, and result (or success [ie, attainment of expected outcome]).19 Overall, despite some interstudy variability, a common concept in many PS definitions is “the use of sedative medications to relieve intolerable suffering from refractory symptoms by a reduction in patient consciousness.”20(p67),21
The most common refractory symptoms requiring sedation are reported in Figure 2. Delirium and dyspnea are fairly unequivocal, frequently present at the end of life, and prognostic for death in a short time.22,23 However, many other symptoms are much harder to characterize. For example, although psychological distress is reported to
be a common reason for PS in several studies,6,9–12 it is seemingly entirely absent in others.5,8
Indeed, PS as a means to treat psychological suffering is particularly
delicate and controversial, presenting some irregularities.
First, it may occur early in the disease
trajectory, not just in the terminal phase. Second, it has little chance
of being alleviated by supportive and/or psychological interventions, so many people should be sedated if their psychological
suffering cannot be controlled. Last, although it tends not to be progressive, such as the suffering caused by physical symptoms, it often
fluctuates and is unpredictable. For all these reasons, PS should be
used with extreme caution, together with regular
multidisciplinary discussions, specialist
psychological support and spiritual assistance, and intermittent or
relief sedation
instead of continuous sedation.24,25 Sedation, as evidenced in this study, may be superficial or deep, continuous or intermittent, gradual or rapid, and primary
or secondary.26 As already noted, the prevalence of PS to control refractory symptoms varies considerably. This could be a reflection of the uncertain definition of the term PS, with some authors focusing on the most extreme continuous deep sedation and others on any type of sedation. The use of different patient case mixes in albeit similar palliative care programs could also be an important factor in this uncertainty.
Sedation is generally used over a short period, and most evidence shows that in the context of specialist palliative care and when correctly used for symptom relief, it is not associated with shortening of life. In this way, from an ethical point of view, the theory of double effect does not apply. Morita et al27 argue this case, reporting a possible negative impact on survival (in which the theory of double effect is rooted) in only 3.9% of sedated patients. Naturally, PS has to be performed with great accuracy, because a voluntary or involuntary abuse of drugs could lead to iatrogenic overdose and acceleration of death. This unwelcome event could have dramatic consequences, especially in nonterminal and respite sedations. For this reason,
the effects of
PS need to be accurately monitored patient by patient, and a correct
approach in decision making and in performing PS is
mandatory. Guidelines and frameworks have recently
been published to help the clinician in this difficult and delicate
area.24,25
Length of survival cannot be considered the only outcome measure of PS; other real outcomes such as ability to control symptoms and probability to survive after discontinuation of PS could be important aspects in this setting. Nevertheless, we decided to focus our review primarily on the former aspect,
being an ethically sensible topic.
Our review reports the data across studies to estimate the length of survival in both sedated and nonsedated patients with more precision than is possible in a single study. However, our review
has several limitations, the most important being the quality of the studies, because randomization is ethically implausible. As a consequence, only prospective cohort and retrospective
studies, not always of good quality, are able to provide the highest level of evidence with the well-known risk of bias. Our quality evaluation of the studies was performed according to Hawker's method, which offers the possibility of reviewing disparate data systematically.4 According to this method, the quality of evidence was quite low: fair to poor in five studies, and fair in six studies.
Other limitations include the heterogeneity of the patient inclusion criteria and of the definitions of PS and differences in clinical setting and type of drug used. Incomplete data on study design, patient population, and criteria for the choice of nonsedation control may hamper interpretation and synthesis of the included studies. These limitations make the presentation of data in a statistical meta-analytic, evidence-based approach almost impossible.
Although the studies included in this review assessed all types of sedation, there were no reports comparing the effects of nonsedation and continuous deep sedation. In each study, there was no significant difference in overall survival between
patients who received sedation or those who did not, although in five studies there was a trend in favor of sedation.
Sedation should be appropriately used for the control of specific symptoms once all other therapeutic alternatives have been considered and found to be ineffective or inapplicable.
As agents to control symptoms, not to shorten life, sedatives should be provided in doses that are titrated against the response to achieve relief of symptoms. Benzodiazepines remain the most favored class of
sedatives in the palliative care world. Midazolam is the most commonly
used drug; it is administered by continuous subcutaneous
infusion, and it has anticonvulsant, muscle
relaxant, and anxiolytic properties. However, the psychotropic drugs
haloperidol,
levomepromazine, and chlorpromazine may be more
appropriate for the specific management of delirium and can be used in combination with benzodiazepines.
In conclusion,
the key drawback of this study is the lack of
evidence from randomized controlled trials, in which patients are
randomly allocated to sedation or nonsedation groups, but
this is an impossible task, because it cannot be
ethically justified. However, our systematic review seems to show that
apart
from observational clinical studies, a certain
level of evidence on the absence of an impact of PS on survival can be affirmed, and PS can therefore be considered as an integral part of the palliative medicine approach to end-of-life care.
© 2012 by American Society of Clinical Oncology
No comments:
Post a Comment