Use of Palliative Oxygen in Cancer Patients

Abstract

Background: Despite the lack of evidence to support the use of palliative oxygen to relieve dyspnea at the end of life, its prescription is widespread and often supported by local and national practice guidelines. Objectives: The objectives of this study were (1) to determine to what extent oxygen prescriptions meet the proposed prescription criteria in our institution, (2) to examine the indication of individual prescriptions in relation to the severity of dyspnea and (3) to review the utilization of opioids in patients receiving palliative oxygen. Methods: Retrospective chart review of cancer patients who were prescribed palliative oxygen between April 2015 and January 2020 through a respiratory home care program in Quebec City, Canada. According to provincial prescription guidelines, palliative oxygen was provided and reimbursed in case of severe hypoxemia (pulse oximetry saturation at rest < 88%) in cancer patients with an estimated prognosis of less than 3 months. Results: 134 patients receiving palliative oxygen were included; 25 (19%) did not fulfill reimbursement criteria. Median survival was 44 days. At initiation of palliative oxygen, 48 patients (36%) had only mild or moderate dyspnea (Medical Research Council dyspnea score 1-3), 26 (19%) did not receive opioids, and 9 (7%) were prescribed palliative oxygen without being dyspneic or receiving opioids. Conclusion: Most prescriptions of palliative oxygen met the proposed prescription criteria in our institution. Half of those who received palliative oxygen were only mildly dyspneic and/or were not receiving opioids at the time of the prescription.

Introduction

Dyspnea at the end of life is a frequent symptom, occurring in 21 to 79% of patients with advanced cancer and is moderate to severe in 10 to 63% of them. ¹ Dyspnea is often multifactorial and difficult to control. It may be distressing, not only for patients but also for their families. ² According to the patient’s performance status and prognosis, its management first includes the identification and treatment of reversible causes. Otherwise, pharmacological interventions to relieve dyspnea including opioids, benzodiazepines and corticosteroids are available. ³ Non pharmacological interventions such as oxygen, breathing training, directing a fan at the patient’s face and walking aids may also be considered.^2,4-7 Pharmacological and non-pharmacological interventions may be applied in combination or in sequence, although no consensus exists as to whichever should be prescribed first.^8,9

The scientific evidence to support the use of oxygen therapy to relieve dyspnea in cancer patients (herein called “palliative oxygen”) is scarce and its prescription is controversial. Two systematic reviews and meta-analyses of interventions for alleviating cancer-related dyspnea included a small number of randomized trials of oxygen, most of which enrolled only non-hypoxemic patients (ie, transcutaneous pulse oximetry saturation [SpO₂] ≥ 90%).^10,11 In 3 of the 6 published randomized trials summarized in these systematic reviews, the intervention was short (<1 hour). Overall, the quality of evidence was deemed as to be poor. Both systematic reviews found no benefit of oxygen on dyspnea. The evidence to support the use of opioids is better than that of oxygen,^12,13 although direct comparison of the 2 interventions from randomized trial is still lacking. Consequently, opioids are still considered by many as the mainstay for managing dyspnea at the end of life.^14,15

Despite the lack of evidence to support the use of palliative oxygen, its prescription is widespread, often supported by local and national practice guidelines, and even reimbursed.^8,16 Our respiratory home care program provides palliative oxygen under specific conditions in patients with advanced cancer. The general objective of our study was to review the practice and the prescriptions of palliative oxygen in our home respiratory care program. More specifically, we wished (1) to determine to what extent oxygen prescriptions meet the proposed prescription criteria in our institution, (2) to examine the indication of individual prescriptions in relation to the severity of dyspnea and (3) to review the utilization of opioids in patients receiving palliative oxygen.

Methods

Setting and Patients

This retrospective chart review took place in the respiratory home care program of the Quebec City area (province of Quebec, Canada) and received approval from the Research Ethics Committee of our institution (CER-IUCPQ-UL: 2019-3184, 21 718). This program is funded by the Quebec universal medical insurance plan and delivers home care (mainly oxygen therapy and related services) to patients with chronic respiratory diseases. According to provincial prescription guidelines, palliative oxygen is provided and reimbursed in case of severe hypoxemia (SpO₂ at rest < 88%) in cancer patients with a prognosis estimated to be less than 3 months. ¹⁷ Dyspnea or previous interventions to relieve it are not required for admission to the program. The vast majority of patients who are prescribed palliative oxygen also receive home care from their Local Community Service Centre (CLSC). Patients who received a prescription of palliative oxygen between April 2015 and January 2020 were included, regardless of their diagnosis. We limited our analysis to those for whom data regarding the severity of dyspnea and current medication were available. Oxygen was always delivered from an electrically-powered oxygen concentrator and was administered through nasal cannula.

Data Extraction

Patients’ medical records and medication charts were searched for the following general information: gender, age, main diagnosis, smoking status, medication (opioids, inhaled bronchodilators, benzodiazepines, diuretics), prescriber (specialist or general practitioner) and whether the prescription was written when the patient was in hospital or at home. The level of dyspnea on the Medical Research Council (MRC) dyspnea scale ¹⁸ (Appendix, Table S1) and the oxygen flow rate (liters per minute) on the day of installation of the oxygen concentrator at home were noted. This information is routinely collected by the visiting respiratory therapist. We also noted whether the oxygen prescription was in accordance with the eligibility criteria of the provincial prescription guidelines. Time of exposure to oxygen (hours per day) was not recorded. Survival time was defined as the time between the installation of palliative oxygen at home (usually within 48 hours of prescription) and death.

Statistical Analysis

Descriptive statistics (means and standard deviations or medians) were used to describe the study population. Overall survival was assessed using Kaplan-Meier estimator. Log-rank test was used to compare survival curves of subgroups of patients. In all analyses, statistical significance was set at the .05 level.

Results

Patients

Between April 2015 and January 2020, 248 patients were prescribed palliative oxygen through our respiratory home care program; 114 patients were excluded because of missing data, leaving 134 patients to contribute to this analysis. Baseline characteristics of these 134 patients are shown in Table 1. Overall, 96% of patients who received a prescription of palliative oxygen had cancer (mostly primary lung cancer). Most prescriptions were written by specialists (pulmonologists). The overall survival curve is shown in Figure 1. Median survival was 44 days. At 3- and 12-months follow-up, 47 (35%) and 15 (12%) of the 134 were still alive, respectively. Baseline characteristics of those who contributed to this analysis and those who were excluded because of missing data pertaining to dyspnea and medication were similar (Appendix, Table S2).

Table 1.

Baseline Characteristics (n = 134).

Female Sex - No. (%)	74 (55)
Age mean - yr	71 ± 10
Active smoking - no./total no. (%) a	13/100 (13)
Dyspnea mean - MRC score	3.7 ± 0.9
Baseline medication - no./total no. (%) a
Opioids	108/134 (80)
Bronchodilators	89/127 (70)
Benzodiazepine	69/121 (57)
Diuretics	38/120 (32)
Diagnosis - no. (%)
Lung cancer	99 (74)
Non-lung cancer	28 (21)
Undetermined	2 (1)
Other	5 (4)
Prescriber - no./total no. (%) a
Specialists	81/133 (61)
General practitioners	52/133 (39)
Prescription context - no./total no. (%) a
Hospital discharge	54/108 (50)
Ambulatory	54/108 (50)
Oxygen flow - liters per minute	2.3 ± 0.9
Met prescription criteria - no. (%)
Yes	76 (57)
No	25 (19)
Undetermined	33 (24)

^aData not available for all patients; Plus–minus values are means ± standard deviations.

Figure 1.

Overall survival. Time to death is from installation of palliative oxygen at home.

Eligibility Criteria

Twenty five of the 134 (19%) patients did not meet eligibility criteria to our program but nevertheless were provided with palliative oxygen. Fourteen were not severely hypoxemic (SpO₂ ≥ 88%), and 8 had a prognosis of more than 3 months; one patients was neither hypoxemic at rest nor had an estimated prognosis of more than 3 months. Two patients had nocturnal hypoxemia alone. The reasons why those who did not meet the prescription criteria received it anyway were not noted. There was no statistical difference in survival between patients who received palliative oxygen in accordance with eligibility criteria and those who did not meet criteria (Appendix, Figure S2). Similarly, we found no difference in survival according to diagnosis (lung cancer vs cancer from other origin vs others), prescriber (specialist vs general practitioner), context of prescription (hospital discharge vs ambulatory clinic) and oxygen flow rate at initiation (≤2 L per minute vs > 2).

Palliative Oxygen vs Severity of Dyspnea and Prescription of Opioids

At initiation of palliative oxygen, 48 patients (36%) had only mild or moderate dyspnea (MRC trade 1 to 3); 26 patients (20%) did not receive opioids (Table 2). Nine patients were prescribed palliative oxygen without being dyspneic or receiving opioids. There was no statistical difference in survival between patients with dyspnea graded 1-3 vs 4-5 and between patients receiving opioids or not (Figure 2).

Table 2.

Palliative Oxygen vs Severity of Dyspnea and Precription of Opioids.

		MRC Dyspnea Score
		1-2-3	4-5
Opioids	Yes	39 (29%)	69 (51%)
	No	9 (7%)	17 (13%)

Figure 2.

Survival according to MRC dyspnea score (left) and opioid prescription (right).

Discussion

We report the use of palliative oxygen in terminally ill patients within a respiratory home care program that is subject to provincial prescription guidelines. We found that most prescriptions were within the boundaries of what is usually recommended to allow reimbursement. The most important finding of our study is that a significant proportion of patients who received palliative oxygen were only mildly dyspneic and/or were not receiving opioids at the time of prescription. A plausible explanation to this situation is that physicians most often rely only on SpO₂ rather than symptoms to prescribe palliative oxygen.

Published systematic reviews and meta-analyses of palliative oxygen in cancer patients have limitations as they mostly include short-term trials in non-hypoxemic patients.^10,11 The largest randomized trial of palliative oxygen in relief of dyspnea in patients with refractory dyspnea is a long-term, multicenter double-blinded randomized controlled trial that compared oxygen 2 L per minute vs ambient air at least 15 hours per day for only 7 days in non-hypoxemic patients (PaO2 ≥ 55 mm Hg) with life-limiting illness and refractory dyspnea. ¹⁹ Cancer patients represented less than 20% of the total study population of 239 patients. A reduction in symptoms was noted in both groups, with no statistically significant difference. Subgroup analyses indicated that the underlying diagnosis did not determine response to palliative oxygen. To add to this evidence, an Australian observational study including 413 terminally ill patients receiving oxygen therapy did not show a significant difference in dyspnea at 1 or 2 weeks; 150 patients (36%) reported an improvement of at least 20% in mean dyspnea scores. ²⁰ Of note, in this study, the use of oxygen by the patient was not specified.

In line with the current evidence, the most recent clinical practice guidelines of the European Society for Medical Oncology (ESMO) for the treatment of dyspnea in advanced cancer patients do not recommend the routine use of palliative oxygen. ⁹ On the contrary, the latest guidelines on the management of dyspnea in advanced cancer published by the American Society of Clinical Oncology (ASCO) recommend standard supplemental oxygen in patients with hypoxemia (saturation is < 90%), especially when dyspnea or respiratory distress are noted. ⁸ In the absence of dyspnea or respiratory distress, the ASCO still recommends the use of oxygen “for oxygenation purpose”, unless vital prognosis is grim.

Another recommendation of the ASCO is the hierarchical approach to introduce interventions in a stepwise manner to include nonpharmacological measures (such as supplemental oxygen) before pharmacological measures (such as opioids). On the contrary, it is the opinion of the ESMO Guidelines Committee that “opioids are the only pharmacologic agents with sufficient evidence in the palliation of dyspnea”. ⁹ Indirect comparisons suggest that opioids are more effective than oxygen to relieve dyspnea.^12,13 Oxygen being expensive and cumbersome, clinicians may therefore prefer introducing low-dose opioids to relieve dyspnea rather than prescribing palliative oxygen.

In order to explain the continuing prescription of palliative oxygen despite lack of evidence, another systematic review of the literature compared patients’, caregivers’ and clinicians’ experiences of palliative oxygen and identified differing perceptions regarding its benefits and burdens. ²¹ Perceived benefits, paucity of evidence, reactive prescribing in response to patients’ and caregivers’ expectations and dependence to oxygen are all factors that perpetuate the prescription of palliative oxygen. The authors emphasized that oxygen use may generate differing goals of therapy for patients and caregivers.

Our finding that almost half of those who received palliative oxygen were only mildly dyspneic at the time of the prescription and/or did not receive opioids is of particular interest. This emphasizes the importance for clinicians of a thorough evaluation even before presenting the option of palliative oxygen to patients and their families. This evaluation should not be limited to the measurement of SpO₂. We would argue that screening for hypoxemia using transcutaneous pulse oximetry in palliative medicine is irrelevant in the absence of significant dyspnea. Pharmacologic and non-pharmacologic alternatives to palliative oxygen should first be investigated. We would disagree with the suggestion that a therapeutic trial of low-dose oxygen be considered in selected non-hypoxemic patients with dyspnea, ⁸ specially when therapy is initiated at home. Our experience is that once an oxygen concentrator is installed in a patient’s home, it is difficult to withdraw the intervention and to recuperate the equipment without inducing further anxiety.

Our study has limitations. The most obvious is that it is a retrospective chart review that relied on information that had not been systematically collected. Consequenlty, missing data was an issue and we cannot exclude that selection bias operated. However, the 114 patients who were excluded because of missing data pertaining to dyspnea and medication had similar characteristics at baseline than those 134 patients who were included (Appendix, Table S2). We interpret this finding as an indication that those who contributed to the analysis are likely representative of the target population.

Conclusion

Use of palliative oxygen is widespread despite the paucity of evidence to support its prescription. The reasons for the continuing utilization of palliative oxygen are numerous and deeply anchored into clinical practice. Further long-term studies to better define the indications of palliative oxygen and its pharmacologic and non-pharmacologic alternatives are needed. In the meantimne, prescription of palliative oxygen should not rely only on SpO₂. The focus should be on the symptom of dyspnea. At the current stage of knowledge, opioids should precede palliative oxygen in the management of the dyspneic and terminally ill patients.

ORCID iD

Yves Lacasse https://orcid.org/0000-0002-4351-5915