Is It Appropriate To Withdraw Antibiotics in Terminal Patients with Cancer with Infection?

Abstract

Background:

Antibiotic administration is frequent in terminal patients with cancer, yet the effects on survival are still under debate.

Objective:

The aim of this study was to examine the status of infection and the benefit/burden of antibiotic administration on the survival of terminal patients with cancer with infection.

Design:

A prospective observational study.

Setting/Subjects:

We studied 799 patients with terminal cancer who were admitted to a palliative care unit in Taiwan between January 2008 and the end of April 2010. Survival was calculated from the first day of admission to the day of death in the palliative care unit or under home care.

Measurements:

A specially designed assessment tool was used daily to evaluate clinical conditions. Afterwards, it was analyzed at different time points in a weekly team meeting. Multivariate Cox proportional hazard analyses were used to examine the benefit/burden of antibiotic administration on survival.

Results:

Four hundred fifty-five patients were diagnosed as having at least one episode of infection after first admission. A total of 295 of the 378 (78.0%) with infection received antibiotic treatment upon admission. Multivariate Cox proportional hazard analyses showed that antibiotic administration was related to improved survival for patients who were still alive 1 week after admission (hazard ratio: 0.66, 95% confidence interfal [CI]: 0.46–0.95). However, antibiotics would be a hazard to patients' survival if used in the time 2 days prior to death (hazard ratio: 1.54, 95% CI: 1.22–1.94).

Conclusions:

The results suggest that with good communication between patients, families, and medical staff, withdrawal of antibiotics should be considered if signs of death appear, in order to avoid unnecessary risks. The possible benefit of prolonged survival should be in line with the goal of care, and also take into account preparing the patient for a dignified death.

Introduction

Terminal patients with cancer are prone to infection due to impaired immunity caused by multiple etiologies.¹ When infection occurs in these patients, physicians may use antibiotics as a form of nonaggressive care. However, administration of antibiotics is often empirical without clear laboratory evidence.² Although different guidelines are available for management of infection in patients with cancer,^3–5 using broad-spectrum antibiotics is still popular near the end of life.⁶ A recent study further reveals that dying patients may create a reservoir of resistant bacteria in intensive care units because they normally receive more antibiotics as treatment in the final days.⁷

The benefit of using antibiotics in terminal patients with cancer has long been a controversial topic. Antibiotics were once considered a life-sustaining treatment, albeit one that did nothing to change the progress of underlying cancer, as cited by the American Medical Association in 1992.⁸ While unable to change the course of malignant diseases, Mirhosseini et al.⁹ suggest using antibiotics to increase the general level of comfort in patients with cancer. However, not all studies agree with this prognosis, especially when antibiotics are administered during the last week of life.^10,11 Antibiotics are found effective in improving specific symptoms associated with infection, such as dysuria caused by urinary tract infection or dysphagia induced by oral candidiasis.¹² When it comes to prolonging survival by using antibiotics, Chen et al.¹³ in Taiwan claim that hospice patients with fever have a significantly longer mean survival and a lower 3-day mortality rate if they receive antibiotic treatment. Lam, et al.¹⁴ also deem that antibiotic therapy according to sensitivity tests is linked to survival more than 14 days in patients with advanced cancer. However, Vitetta et al.¹⁵ argue that bacterial infection is not significantly associated with shortened survival in hospice patients. Oneschuk, et al. also consider antibiotics futile in the last week of life, although their study does not address life prolongation.¹⁶ The effects of antibiotics on survival are still under debate. The decision-making process for using antibiotics depends on good communication between patients, families, and physicians.¹⁷

We presumed that antibiotic treatment, in the case of good communication between patients, families, and medical staff, would be futile in terminal patients with cancer at their last days of life. Therefore, we conducted this prospective observational study to examine the impact of using antibiotics on the survival of patients with advanced cancer. Infection rates, status of antibiotic utilization, and ethical acceptability by our medical staff were also reviewed. The results provide suggestions for the reasonable use of antibiotics at the end of life.

Patients and Methods

Patients

During the study period from January 2008 to the end of April 2010, 844 consecutive patients with advanced cancer were admitted to the palliative care unit of National Taiwan University Hospital. Members of the admission committee performed an initial assessment to identify patients with cancer who were not responsive to curative treatment. The design of this study and selection of patients were approved both by the ethical committee in the hospital (Registry number NTUH200705056R) and by the National Science Council (Registry number 96-2314-B-002-170-MY3) in Taiwan. By the end of the study, 8 patients were still alive, and 37patients had died after transferring to other units of the hospital. After excluding these 45 patients, 799 patients were studied and enrolled in analyses.

Instrument

The assessment tool used was a recording form specially designed by the investigators after careful analysis of the literature in this field. The form was tested for content validity with a panel consisting of physicians, a nursing supervisor, and senior nurses. All were experienced in caring for patients with terminal cancer. In addition, we conducted a pilot study in the same unit for 1 month prior to the formal study. The pilot study further confirmed the content validity and the applicability of the instrument. The instrument recorded demographic data (age and gender), clinical conditions related to cancer (primary cancer sites, metastases, history of past treatment, survival time, outcome), status of infection (infection sites, common symptoms of infection, evidence of bacterial culture, blood test results), the antibiotics being used, the effects perceived by conscious patients with cancer (satisfactory, not changed, or unsatisfactory following the decision to administer or withdraw antibiotic treatment), and the opinions of our medical staff regarding the ethical acceptability of using antibiotics (appropriate, acceptable, or inappropriate following the decision to administer or withdraw antibiotic treatment).

Methods

Infection sites, diagnosis of infection, and the antibiotics used were recorded every day by the same primary care physicians. Since symptoms of infection may be obscure in patients with advanced cancer,^18,19 the primary care physicians made the diagnosis of infection based on microbe-related symptoms, with the assistance of laboratory or imaging reports when available. Data at different time points (admission, 1 week after admission, 2 days before death) were discussed and then analyzed in the 120 weekly team meetings held during the study period. All 799 patients enrolled were completely assessed. Furthermore, the same weekly team meetings investigated the conflicts of opinion regarding antibiotic administration between patients, families, and medical staff on the basis of moral discussions. Survival was calculated from the first day of admission to the day of death in the palliative care unit or under home care.

When terminal patients with cancer were admitted to the palliative care unit, antibiotics used in the beginning were usually prescribed by the referring physicians. After admission to the palliative care unit, however, antibiotic administration was based on available laboratory/imaging data and the goal of care. The decision to use antibiotics 1 week after admission was always made after discussion between patients, families, and medical staff. This cooperative decision-making process marked the effectiveness of good communication.

Statistical analysis

SPSS software version 16.0.2 (SPSS Inc., Chicago, IL) was used for data processing and statistical analysis. Frequency was used to describe the demographic data as well as its variables. Univariate Cox proportional hazard analyses were performed by fitting the following factors into the model: age, gender, administration of antibiotics, primary cancer sites, common symptoms of infection (fever, fatigue, consciousness change, delirium, etc.), and common laboratory findings of infection (increased white blood cells, pyuria, positive bacteria culture in blood, elevated level of C-reactive protein, etc.). Factors with statistical significance in the univariate analyses were then entered into multivariate Cox proportional hazard analyses for further investigation. The statistical significance was set as p<0.05.

Results

The primary cancer sites in these patients were the liver (21.5%), lung (16.6%), and colon/rectum (12.1%). Of the 799 patients, 631 (79.0%) already had metastases. Nearly one-half of the patients (43.3%) died within 1 week after admission, and the mean survival was 18.54±40.66 days. The demographic characteristics are displayed in Table 1.

Table 1.

Demographic Characteristics of Patients

Variables	n	%
Gender
Male	462	57.8
Female	337	42.2
Age group
≤18	2	0.3
19–35	14	1.8
36–50	101	12.6
51–64	253	31.7
≥65	429	53.7
Mean±SD	65.22±13.92
Primary sites of cancer
Liver	172	21.5
Lung	133	16.6
Colorectal	97	12.1
Others	94	11.8
Stomach	58	7.3
Pancreas	49	6.1
Head and neck	47	5.9
Breast	26	3.3
Cervical/uterine	19	2.4
Kidney	18	2.3
Esophagus	17	2.1
Bladder	16	2.0
Brain	15	1.9
Prostate	13	1.6
Uncertain	12	1.5
Skin	7	0.9
Leukemia	6	0.8
Outcome
Death at hospital	680	85.1
Death under home care	55	6.9
Discharge just before death	64	8.0
Survival days
≤3	163	20.4
4–7	183	22.9
8–30	358	44.8
≥31	95	11.9
Mean±SD	18.54±40.66
Total	799	100.0

SD, standard deviation.

Table 2 shows the status of infection and antibiotic administration in terminal patients with cancer. Four hundred fifty-five patients were diagnosed as having infection during hospitalization. Three hundred seventy-eight patients (83.1%) had infection at the time of admission and the leading infection sites were the respiratory system (52.6%), hepato-biliary-pancreatic system (20.4%), and urinary system (19.0%). These three systems were also the major infection sites from admission to 2 days before death. The infection rates were 73.6% and 77.4% at 1 week after admission and 2 days before death, respectively. It was worth mentioning that approximately 80% of the patients with infection did not have fever and had been assessed as having poor immune function in the opinion of the primary care physicians. The percentage of patients receiving antibiotic administration was approximately the same at admission (78.0%) and 1 week after admission (75.8%), but decreased to 59.1% 2 days before death. As far as the types of antibiotics involved, intravenous antibiotics were the most frequently administered type at all time points of the survey, followed by oral antibiotics. Externally applied antibiotics were seldom used.

Table 2.

Frequencies, Antibiotic Administration, and Infection Sites in Terminal Patients with Cancer

	Admission		One week after admission		Two days before death
	No.	%	No.	%	No.	%
Infection status
Total	455	100.0	242^a	100.0	411^b	100.0
No infection	77	16.9	64	26.4	93	22.6
With infection	378	83.1	178	73.6	318	77.4
Antibiotic administration
Total	378	100.0	178	100.0	318	100.0
No administration	83	22.0	43	24.2	130	40.9
With administration	295	78.0	135	75.8	188	59.1
Infection sites (multiple choices)
Mouth/throat	12	3.2	4	2.2	5	1.5
Cancer wound	22	5.8	14	7.9	21	6.6
Brain	5	1.3	1	0.6	3	0.9
Respiratory system	199	52.6	90	50.6	201	63.2
Gastrointestinal system	63	16.7	26	14.6	43	13.5
Liver/gallbladder/pancreas	77	20.4	36	20.2	54	17.0
Urinary system	72	19.0	42	23.6	69	21.7
Tubes	11	2.9	4	2.2	10	3.1
Others	37	9.8	22	12.4	32	10.1
Total patients with infection	378	100.0	178	100.0	318	100.0

Only 242 patients had survival longer than 1 week.

Only 411 patients had survival longer than 2 days.

Table 3 demonstrates the effects and the moral acceptability of antibiotic administration. Two hundred forty-four of the 282 patients (86.5%) rated the decision to administer or withdraw antibiotics on admission—as evaluated by our medical staff from the ethical viewpoint—as satisfactory. Meanwhile, antibiotic treatment was recognized as appropriate by our medical staff in the weekly ethical discussion for the majority (82.0%) of the patients receiving it. Satisfaction toward using antibiotics or not was always close to 90% in the patients of our study, but the medical staff's acceptance of antibiotic administration declined slightly as the time of death drew near. This might imply possible conflicts among patients, families, and medical staff when the benefits of using antibiotics become unclear.

Table 3.

Effects Perceived by Conscious Patients Subjectively and Ethical Acceptability of Medical Staff Regarding Antibiotic Administration

	Admission		One week after admission		Two days before death
	No.	%	No.	%	No.	%
Effects perceived regarding antibiotic administration by conscious cancer patients with infection under palliative care
Unsatisfactory	2	0.7	0	0.0	2	1.0
Not changed	36	12.8	14	10.2	23	11.2
Satisfactory	244	86.5	123	89.8	180	87.8
Conscious patients among patients with infection	282	100.0	137	100.0	205	100.0
Total patients with infection	378		178		318
Medical staff's acceptance regarding antibiotic administration in cancer patients with infection under palliative care from ethical viewpoint
Inappropriate	12	3.2	8	4.5	15	4.7
Acceptable	56	14.8	16	9.0	46	14.5
Appropriate	310	82.0	154	86.5	257	80.8
Total	378	100.0	178	100.0	318	100.0

Many families and medical staff think that withdrawing antibiotic treatment in terminal patients with cancer may shorten the patients' lives, prompting us to investigate the associations between antibiotics and patients' survival. The report regarding multivariate Cox proportional hazard analyses are shown in Table 4, and only those variables reaching statistical significance are displayed. At the time of admission, consciousness change (hazard ratio: 1.45, 95% [confidence interval] CI: 1.12–1.89), dyspnea (hazard ratio: 1.35, 95% CI: 1.10–1.67), abnormal liver function (hazard ratio: 1.31, 95% CI: 1.05–1.63), and abnormal kidney function (hazard ratio: 1.42, 95% CI: 1.13–1.80) indicated shorter survival in terminal patients with cancer. Use of antibiotics on admission did not reach a significant level (hazard ratio: 0.82, 95% CI: 0.64–1.05). If the patients were still alive 1 week after admission, fever (hazard ratio: 1.46, 95% CI: 1.06–2.02) as well as abnormal kidney function (hazard ratio: 1.58, 95% CI: 1.01–2.46) remained two significant hazards to patients' lives. Antibiotic administration (hazard ratio: 0.66, 95% CI: 0.46–0.95) was found to prolong survival for patients with hospital stay longer than 1 week. However, patients' survival would decrease if antibiotics were administered 2 days prior to death (hazard ratio: 1.54, 95% CI: 1.22–1.94).

Table 4.

Associations of Survival with Infection-Related Observations in Infected Patients with Cancer under Palliative Care by Multivariate Cox Proportional Hazard Analyses ^a

Time	Variable	Hazard ratio (95% CI)	p value
Admission	Consciousness change	1.451 (1.118–1.885)	0.005^b
(n=378)	Delirium	0.997 (0.785–1.267)	0.982
	Dyspnea	1.354 (1.098–1.669)	0.005^b
	Abnormal liver function	1.309 (1.054–1.627)	0.015^b
	Abnormal kidney function	1.424 (1.128–1.799)	0.003^b
	Antibiotic administration	0.818 (0.636–1.052)	0.117
One week after admission	Fever	1.461 (1.056–2.021)	0.022^b
(n=178)	Consciousness change	1.245 (0.877–1.769)	0.220
	Abnormal kidney function	1.576 (1.011–2.458)	0.045^b
	Antibiotic administration	0.664 (0.464–0.951)	0.025^b
Two days before death	Abdominal pain	1.238 (0.969–1.583)	0.088
(n=318)	Increased WBC with abnormal differential count	1.084 (0.813–1.447)	0.583
	Increased urine WBC	1.267 (0.825–1.947)	0.280
	Abnormal liver function	1.166 (0.843–1.612)	0.354
	Abnormal kidney function	1.190 (0.837–1.691)	0.332
	Antibiotic administration	1.541 (1.222–1.943)	0.000^b

Only those variables reaching statistical significance in univariate Cox proportional hazard analyses were included in the multivariate Cox proportional hazard model.

p<0.05 by multivariate Cox proportional hazard analyses.

CI, confidence interval; WBC, white blood cell (count).

Discussion

Antibiotic administration has been regarded as life-sustaining treatment to prolong life in the opinions of some doctors and families, and thus is common in hospitalized patients during the last days of life, even in palliative care units.^1,2,10,20 Previous studies only discussed the types of infection, the patterns of bacteria isolated, and the categories of antibiotics used in hospice patients.^6,15,21 Our study is unique because it prospectively explores the impact of antibiotics on the survival of terminal patients with cancer when the collaborative decision-making process plays a role. If more studies are conducted regarding the reasonable use of antibiotics, it will lead to better care and cost effectiveness in both palliative care units and general wards.

In our study, the infection rates were similar when patients were approaching death, but the percentage of antibiotics used was lower at 2 days before death (Table 2). Meanwhile, the satisfaction perceived by conscious patients and the ethical acceptability as rated by our medical staff were always higher than 80% (Table 3). These findings reflect good collaboration between the patients, their families, and the palliative care team. In Taiwan, patients facing imminent death often encounter the dilemma of antibiotic administration, which leads to reduced quality of life because both families and medical staff regard antibiotics as “essential” in the treatment of infection.²² However, after detailed explanation of inevitable multiorgan failure, and in keeping with the study of White et al.,²³ many patients wish to withdraw or limit antibiotics to symptom relief only. Without discussing antibiotics, patients accept more easily the holistic care a good palliative care unit has to offer. As Ford et al.¹⁷ point out, the decision to use anti-infective agents during the last days of life requires good information and process. When facing the dilemma of antibiotics, physicians should always take into account benefits and burdens as they apply to each individual situation.²⁴ This was achieved in our study by ensuring good doctor–patient communication when cancer progressed irreversibly.

Concerning the ethical aspect, withholding antibiotics in terminal patients has been viewed as a measure of palliative care under specific situations.^25,26 In 1997, The Supreme Court of the United States ruled that withdrawal of life support was different from assisted suicide.²⁷ According to these viewpoints, discontinuation of life support is ethically acceptable if the intention behind doing so is to stop futile treatment. In our study, the findings further suggest that antibiotic treatment in the last two days of life may be harmful to the patients, justifying the ethical consideration to withdraw antibiotics near the time of death. Other ethical concerns to withdrawing antibiotics include limited effectiveness in controlling symptoms,²³ inconsistency with the hospice goal of care, increased chance of resistant pathogens, and possible delay of the dying process.¹⁷

In terms of the legal aspect, the Congress of the United States passed the Patient Self-Determination Act in 1990.²⁸ By promoting advance directives, the Act helps patients or a surrogate decision-maker refuse unwanted medical treatments.²⁹ Looking back at the situation in Taiwan, the Natural Death Act in June 2000 also gives terminal patients or their families the rights to avoid unnecessary treatments if they only prolong the dying stage.30 This Act not only represents the respect for autonomy, but also increases the need for a holistic vision. In Taiwan, the study of Chiu et al.²² found that having a positive attitude toward the Natural Death Act helped decrease the ethical dilemma. Good continuous education also increases understanding of the Act in health care providers.³¹ Hence, we can conclude that withdrawing antibiotics after a thorough discussion is not a violation of the laws, but an ethical act of clinical professionalism.

Since the findings in our study indicate the benefits of withdrawing antibiotics near the time of death, the diagnosis of imminent death becomes essential because it helps determine appropriate treatments for terminal patients.³² Although there is still no accurate method to determine exactly when death will occur, some symptoms and signs have been suggested as indicative of the terminal stages of the dying process. For example, the death rattle can predict that a patient will die within 2 days.³³ Other signs often observable in the dying phase of patients with cancer include: bedbound, semicomatose, being able to take only sips of fluid, and no longer able to take oral tablets.³⁴ In another study conducted by Morita et al.,³⁵ 84% and 92% of patients showed deterioration of consciousness 24 and 6 hours before death, respectively. The evidence above shows that some “signs of death” are important indicators of imminent death. Thus, we suggest withdrawing antibiotics in patients with cancer displaying a significant number of the aforementioned signs of death after careful evaluation by physicians. However, in adherence to guidelines, reasonable use of antibiotics is still advised in terminal patients with cancer before such signs of death appear.

Several limitations should be mentioned in relation to our study. First, only terminal patients with cancer who died in the palliative care unit of a single medical center were included in our study and this might have limited the generalizability of the results. Nevertheless, the study hospital is a tertiary hospital, which accommodates a variety of patients from different regions of Taiwan. Second, symptoms of infection in terminal patients with cancer can sometimes be indistinct or even absent. For example, patients with infection may not always have fever. However, the same primary care physicians assessed every patient's condition in our study, and daily discussion was held to minimize misdiagnosis. Third, we did not use an objective questionnaire to assess the conflicts between families and medical staff; instead, we surveyed ethical acceptability among medical staff and satisfaction among patients toward using antibiotics. We believe this close observation from the aspect of care can show our good mutual communication. Finally, our study is observational due to ethical considerations, and the study population is a mixed group ranging from imminently dying patients to temporarily stable patients. Further studies may categorize patients according to different types of cancer and infection to identify those who can derive the most benefit from antibiotic treatment.

In conclusion, using antibiotics in terminal patients with cancer should be carefully evaluated before signs of death appear. However, if patients show signs of death, withdrawal of antibiotics should be considered in the case of good communication between patients, families, and medical staff to avoid unnecessary risks. The possible benefit of prolonged survival should be in line with the goal of care and should take into account the preparation for a dignified death.

Footnotes

Acknowledgments

An-Hsuan Chih and Tai-Yuan Chiu had full access to all the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis. Study concept and design: Tai-Yuan Chiu.

Acquisition of data: Long-Teng, Lee, Shao-Yi Cheng, Chien-An Yao.

Analysis and interpretation of data: An-Hsuan Chih, Tai-Yuan Chiu.

Drafting of the manuscript: An-Hsuan Chih. Critical revision of the manuscript for important intellectual content: Ching-Yu Chen, Tai-Yuan Chiu. Statistical analysis: An-Hsuan Chih, Tai-Yuan Chiu. Administrative, technical, or material support: Tai-Yuan Chiu.

Study supervision: Tai-Yuan Chiu. Research support: National Science Council, Taiwan

Author Disclosure Statement

No competing financial interests exist.

References

Pereira

, Watanabe

, Wolch

. A retrospective review of the frequency of infections and patterns of antibiotic utilization on a palliative care unit. J Pain Symptom Manage, 1998; 16:374–381.

Ahronheim

, Morrison

, Baskin

, Morris

, Meier

. Treatment of the dying in the acute care hospital. Advanced dementia and metastatic cancer. Arch Intern Med, 1996; 156:2094–2100.

Freifeld

, Bow

, Sepkowitz

et al. Clinical practice guideline for the use of antimicrobial agents in neutropenic patients with cancer: 2010 update by the Infectious Diseases Society of America. Clin Infect Dis, 2011; 52:e56–93.

Rolston

. The use of new and better antibiotics for bacterial infections in patients with leukemia. Clin Lymphoma Myeloma, 2009; 9,Suppl 3:S357–S363.

Hicheri

, Toma

, Maury

et al. Updated guidelines for managing fungal diseases in hematology patients. Expert Rev Anti Infect Ther, 2010; 8:1049–1060.

Chun

, Rodgers

, Vitale

, Collins

, Malani

. Antimicrobial use among patients receiving palliative care consultation. Am J Hosp Palliat Care, 2010; 274:261–265.

Levin

, Simor

, Moses

, Sprung

. End-of-life treatment and bacterial antibiotic resistance: A potential association. Chest, 2010; 138:588–594.

Council on Ethical and Judicial Affairs, American Medical Association. Decisions near the end of life. JAMA, 1992; 267:2229–2233.

Mirhosseini

, Oneschuk

, Hunter

et al. The role of antibiotics in the management of infection-related symptoms in advanced cancer patients. J Palliat Care, 2006; 22:69–74.

10.

, Kim

et al. Antibiotic use during the last days of life in cancer patients. Eur J Cancer Care (Engl), 2006; 15:74–79.

11.

Nakagawa

, Toya

, Okamoto

et al. Can anti-infective drugs improve the infection-related symptoms of patients with cancer during the terminal stages of their lives? J Palliat Med, 2010; 13:535–540.

12.

Enck

. Antibiotic use in end-of-life care: a soft line? Am J Hosp Palliat Care, 2010; 27:237–238.

13.

Chen

, Chou

, Hsu

et al. Antibiotic prescription for fever episodes in hospice patients. Support Care Cancer, 2002; 10:538–541.

14.

Lam

, Chan

, Tse

, Leung

. Retrospective analysis of antibiotic use and survival in advanced cancer patients with infections. J Pain Symptom Manage, 2005; 30:536–543.

15.

Vitetta

, Kenner

, Sali

. Bacterial infections in terminally ill hospice patients. J Pain Symptom Manage, 2000; 20:326–334.

16.

Oneschuk

, Fainsinger

, Demoissac

. Antibiotic use in the last week of life in three different palliative care settings. J Palliat Care., 2002; 18:25–28.

17.

Ford

, Fraser

, Davis

, Kodish

. Anti-infective therapy at the end of life: Ethical decision-making in hospice-eligible patients. Bioethics, 2005; 19:379–392.

18.

Homsi

, Walsh

, Panta

et al. Infectious complications of advanced cancer. Support Care Cancer, 2000; 8:487–492.

19.

Bodey

. Infection in cancer patients. A continuing association. Am J Med, 1986; 81:11–26.

20.

Goodlin

, Winzelberg

, Teno

, Whedon

, Lynn

. Death in the hospital. Arch Intern Med, 1998; 158:1570–1572.

21.

Reinbolt

, Shenk

, White

, Navari

. Symptomatic treatment of infections in patients with advanced cancer receiving hospice care. J Pain Symptom Manage, 2005; 30:175–182.

22.

Chiu

, Hu

, Huang

, Yao

, Chen

. Prevailing ethical dilemmas in terminal care for patients with cancer in Taiwan. J Clin Oncol, 2009; 27:3964–3968.

23.

White

, Kuhlenschmidt

, Vancura

, Navari

. Antimicrobial use in patients with advanced cancer receiving hospice care. J Pain Symptom Manage, 2003; 25:438–443.

24.

Nagy-Agren

, Haley

. Management of infections in palliative care patients with advanced cancer. J Pain Symptom Manage, 2002; 24:64–70.

25.

Fabiszewski

, Volicer

. Effect of antibiotic treatment on outcome of fevers in institutionalized Alzheimer patients. JAMA, 1990; 263:3168–3172.

26.

Gruzalski

. Nontreatment of fever. N Engl J Med, 1979; 301:784–785.

27.

Vacco v Quill, 521 US 793. 1997.

28.

Omnibus Budget Reconciliation Act of 1990Public law no. 101-508).

29.

Luce

, White

. A history of ethics and law in the intensive care unit. Crit Care Clin, 2009; 25:221–237.

30.

Natural Death Act in Taiwan (Hospice-Palliative Care Act): Promulgated on June 7, 2000 by President's Order Hua Chung (1) Yi Tze No. 8900135080.

31.

Tsai

, Lee

, Lai

et al. Practical effects of educating nurses on the Natural Death Act. Support Care Cancer, 2005; 13:232–238.

32.

Sykes

. End of life issues. Eur J Cancer, 2008; 44:1157–1162.

33.

Wildiers

, Menten

. Death rattle: Prevalence, prevention and treatment. J Pain Symptom Manage, 2002; 23:310–317.

34.

Ellershaw

, Ward

. Care of the dying patient: The last hours or days of life. BMJ, 2003; 326:30–34.

35.

Morita

, Ichiki

, Tsunoda

, Inoue

, Chihara

. A prospective study on the dying process in terminally ill cancer patients. Am J Hosp Palliat Care, 1998; 15:217–222.