This may be the author’s version of a work that was submitted/acceptedfor publication in the following source:

Ekstrom, Magnus, Allingham, Samuel, Eagar, Kathy, Yates, Patsy, John-son, Claire, & Currow, David(2016)Breathlessness during the last week of life in palliative care: an Australianprospective, longitudinal study.Journal of Pain and Symptom Management, 51(5), pp. 816-823.

This file was downloaded from: https://eprints.qut.edu.au/96095/

c© Consult author(s) regarding copyright matters

This work is covered by copyright. Unless the document is being made available under aCreative Commons Licence, you must assume that re-use is limited to personal use andthat permission from the copyright owner must be obtained for all other uses. If the docu-ment is available under a Creative Commons License (or other specified license) then referto the Licence for details of permitted re-use. It is a condition of access that users recog-nise and abide by the legal requirements associated with these rights. If you believe thatthis work infringes copyright please provide details by email to qut.copyright@qut.edu.au

License: Creative Commons: Attribution-Noncommercial-No DerivativeWorks 2.5

Notice: Please note that this document may not be the Version of Record(i.e. published version) of the work. Author manuscript versions (as Sub-mitted for peer review or as Accepted for publication after peer review) canbe identified by an absence of publisher branding and/or typeset appear-ance. If there is any doubt, please refer to the published source.

https://doi.org/10.1016/j.jpainsymman.2015.12.311

1

Breathlessness During the Last Week of Life in Palliative Care: an

Australian prospective, longitudinal study

Magnus Ekström, MD, PhD1,2

; Samuel F Allingham, BMath(Hons)3; Kathy Eagar, MA, PhD,

FAFRM (Hon)3; Patsy Yates, PhD, RN, FAAN, FACN

4; Claire Johnson,

PhD, RN

5; David C

Currow, B Med, MPH, PhD, FRACP, FAHMS1.

1. Discipline, Palliative and Supportive Services, Flinders University, Adelaide, Australia

2. Department of Clinical Sciences, Division of Respiratory Medicine & Allergology, Lund

University, Lund, Sweden

3. Palliative Care Outcomes Collaboration, Australian Health Services Research Institute,

University of Wollongong, Wollongong, NSW, Australia

4. School of Nursing, Institute of Health and Biomedical Innovation, Queensland University

of Technology, Brisbane, Australia

5. The Cancer and Palliative Care Research and Evaluation Unit, School of Surgery, The

University of Western Australia, Perth, Australia

Corresponding author: Magnus Ekström, Discipline of Palliative and Supportive Services,

Flinders University, Adelaide, Australia; Email: pmekstrom@gmail.com; Telephone:

+61447485097.

Funding: PCOC is funded by the Australian Government Department of Health. M.E. was

supported by unrestricted grants from The Swedish Society of Medicine, the Swedish

Respiratory Society, the Swedish Heart-Lung Foundation, the Scientific Committee of

Blekinge County Council, and the Wera and Emil Cornell Foundation.

Running head: Breathlessness in the last week of life

Number of tables: 4; Number of figures: 2; Number of references: 4037

Word count for body of text: 32,224581

Formatted: Not Highlight

Formatted: Not Highlight

2

Word count for abstract: 240

ABSTRACT

Context: Breathlessness is a major cause of suffering and distress and little is known about

the trajectory of breathlessness near death.

Objectives: To determine the trajectory and clinical-demographical factors associated with

breathlessness in the last week of life in specialist palliative care.

Methods: Prospective, longitudinal cohort study using national data in specialist palliative

care from the Australian Palliative Care Outcomes Collaboration (PCOC). We included

patients in PCOC who died between 1 July 2013 and 30 June 2014 with at least one

measurement of breathlessness on a 0−10 numerical rating scale (NRS) in the week before

death. The trajectory and factors associated with breathlessness were analyzed using

multivariate random effects linear regression.

Results: A total 12,778 patients from 87 services (33,404 data points) were analyzed. The

average observed breathlessness was 2.1 points and remained constant over time. Thirty-five

percent reported moderate to severe distress (NRS ≥ 4) at some time in their last week.

Factors associated with higher breathlessness were younger age, male gender,

cardiopulmonary involvement, concurrent fatigue, nausea, pain, sleeping problems, higher

Australia-modified Karnofsky Performance Status, and clinical instability in the multivariate

analysis. Respiratory failure showed the largest association (mean adjusted difference 3.1

points; 95% confidence interval, 2.8 to 3.4).

Conclusion: Although breathlessness has been reported to worsen in the last months, the

mean severity remained stable in the final week of life. In specialized palliative care, one in

three people experienced significant breathlessness especially in respiratory disease.

3

Keywords: Dyspnea; mortality; palliative care; terminal care; respiratory insufficiency; risk

factors; cohort studies.

4

INTRODUCTION

More than 50% of people experience breathlessness at the end of life.(1) These people have a

range of life-limiting illnesses, not limited to cardio-respiratory diseases. Many people at the

end of life have several clinical conditions that contribute to progressive breathlessness.(1)

The prevalence and severity of breathlessness at the end of life varies widely between

individuals with apparently similar disease processes.(2)

Breathlessness is a feared symptom at the end of life, especially in people with respiratory

disease.(3) Breathlessness impairs patients’ quality of life, wellbeing, and function.(4, 5)

For patients and their caregivers, it is a symptom that continuously reminds people of

impending death.(6) Breathlessness have been reported to worsen in people with advanced

disease especially in the last months of life despite evidence-based, symptomatic

treatments.(4, 5, 7) People with worsening breathlessness particularly fear that they may

suffocate.(7)

Half of all caregivers perceive that the person for whom they were caring was

‘uncomfortable’ or ‘very uncomfortable’ in the last two weeks of life.(8)

Little is known about breathlessness in the last week, including any factors that may help to

predict higher levels of symptom prevalence or distress.(9, 10) Understanding the temporal

patterns of breathlessness at the end of life and clinical-demographical factors associated with

breathlessness could help to improve the quality of care that is offered.

Previous studies have focused on breathlessness over longer periods, often with few observations in

the last hours or days of life. From a large, representative, prospectively collected longitudinal

database, are there factors that a clinician could identify that put people at higher risk of more severe

breathlessness at the end of life? Such knowledge would be important to ensure optimized symptom

control.

5

The aim of this study was to describe the patterns of breathlessness by level of distress and

underlying life-limiting illness in the last week of life and to identify any sub-populations who

may be at particular risk of more severe breathlessness. The null hypothesis was that there are

no identifiable factors associated with worse breathlessness in the last week of life.

METHODS

Study design and population

This was a prospective, longitudinal cohort study using data from the Australian Palliative

Care Outcomes Collaboration (PCOC).(11) PCOC is funded by the Australian Government

Department of Health and collects point-of-care data in specialist palliative care services

across Australia including inpatient and community-based care.(11) Data were included from

all participating services that routinely collected symptom severity data. The database has

been used in previous reports.(11, 12)

Inclusion criteria were patients cared for by services registered in PCOC, dying between 1

July 2013 and 30 June 2014, who had at least one breathlessness measurement during the

final seven days of life.

Data collection

Services participating in PCOC (inpatient, consultative, and community settings) provided

data for each patient’s demographic characteristics, underlying disease, setting of care, and

clinical assessments at each change in the Palliative Care Phase (herein ‘phase’), with services

assessing patients each day they are seen.

6

Phase includes four clinically relevant categories that describe each patient’s palliative care

trajectory: stable; unstable; deteriorating; and terminal. The criteria for the start and end of

each phase were recently validated and updated, and the latest version were used in this

study.(12) A new phase was assigned whenever a clinical change required patient/family

reassessment or modification of the care plan.(11, 12)

The standardized assessment included functional status (Australia-modified Karnofsky

Performance Status; AKPS) (13), and distress from each of seven symptoms using the

Symptoms Assessment Scale (SAS): appetite problems; bowel problems; breathing problems;

fatigue; nausea; pain; and difficulty sleeping.(14) SAS measures the severity of the distress

from each symptom over the previous 24 hours on a numerical rating scale (NRS) between 0

(“absent or no distress”) and 10 (“worst possible distress”).(14) Scores were by self-report if

possible or by proxy.(15, 16) Some symptom assessments (6%) were recorded using the

Edmonton Symptom Assessment System (ESAS; 0-10 NRS of severity) due to local routines.

ESAS assessments were included as a sensitivity analysis including/excluding them yielded

consistent findings.

Ethical considerations

The PCOC was approved by the Human Research Ethics Committee of the University of

Wollongong (approval ID: HE06/045). As routine clinical data were collected, separate

consent was not required.

Statistical analyses

7

Data were tabulated using frequencies and percentages for categorical variables, mean with

SD, and median with range or interquartile range for continuous variables with normal and

skewed distribution, respectively.

The primary outcome was distress from breathlessness during the seven days before death.

All measurements for all patients were included in the analysis. Patients with only one

measurement were included as they contributed to the estimation of the average

breathlessness level and association for factors that vary between patients. No data were

imputed.

Breathlessness was analyzed in two ways: as average scores and as percentages of the distress

categories: none (NRS 0); mild (NRS 1 to 3); moderate (NRS 4 to 7); and severe (NRS 8 to

10). (9, 17) Breathlessness over the last seven days of life was explored graphically in relation

to age, gender, underlying disease (malignant vs. non-malignant), lung disease (defined as

chronic obstructive pulmonary disease [COPD] or lung cancer), setting of care (hospital vs.

community), phase, and AKPS.

Associations with breathlessness (0-10 NRS) were estimated using uni- and multivariate

random effects linear regression, accounting for repeated measurements.(18) Variables to

include in the final model were selected based on subject matter knowledge (2, 9, 16, 19) and

by exploring the associations for potential factors with breathlessness and covariates in the

model. Time-varying variables were AKPS, symptoms, and phase. Days before death, age,

and gender were included in all multivariate models. At the final stage, all initial variables

were entered simultaneously to check that the model estimates were robust. The final model

included the variables: day before death (0 to 7); male gender (yes/no); age (continuous);

primary diagnosis (lung cancer, other cancers, cardiovascular disease, respiratory failure, and

other non-malignant disease); AKPS (continuous, 10 to 100); nausea and pain (0−10 NRS);

8

and phase (stable, unstable, deteriorating, or terminal). For time-varying covariates, we

estimated the association for differences between patients by regressing on each patient’s

covariate mean, and the association within patients (longitudinal change between

measurements) by centering each measurement on the covariate mean for each patient.

We calculated 95% confidence intervals (CIs) for all estimates. Statistical analyses were

performed with Stata version 12.1 (StataCorp LP; College Station, TX, USA; 2013).

RESULTS

We included 12,778 patients from 87 palliative care services across Australia, providing

33,404 data points during the last week of life. Patient characteristics are shown in Table 1.

The mean age was 73 (standard deviation [SD], 14) years and 54% were men. Most patients

had cancer (78%) and the most frequent diagnosis was lung cancer (17%). There were 22,057

measurements of breathlessness, a mean 1.7 (SD, 0.9) measures per patient (median 1; range

1 to 8) in the last week. Forty-nine percent of patients had two or more breathlessness

measurements during the last seven days. The total number of breathlessness measurements

per day is shown in Table 2.

Reported breathlessness

The mean reported breathlessness was 2.1 points on a 0-10 NRS (median 0; interquartile

range 0 to 4) in the last week of life. Breathlessness varied markedly between patients (SD,

2.7), and also within patients over time (SD, 1.3). Fifty-four percent of people reported

breathlessness (NRS ≥1) and 35% had moderate to severe levels (NRS ≥ 4) at some time

during the last week. The observed mean breathlessness from the beginning of the last week

9

of life did not change as death approached (Figure 1), and the time-trends were similar for the

subgroups based on: age (above vs. below 65 years), gender, primary diagnosis (lung cancer,

other cancer, respiratory failure, cardiovascular disease, or other non-malignant disease), and

care setting.

Factors associated with breathlessness

Table 3 shows factors associated with breathlessness during the last week of life in the

multivariate model (n = 9,759 patients with non-missing data on all variables; 13,171

observations). The average reported breathlessness was higher closer to death when adjusted

for the other variables in the model. Higher breathlessness was also independently associated

with male gender; younger age; a primary diagnosis lung cancer, respiratory failure, or other

non-malignant disease; higher functional status, concurrent nausea and pain; and with

unstable or deteriorating clinical phase. The factor associated with the largest difference in

breathlessness was respiratory failure, with a mean increase of 3.1 (95% CI, 2.8 to 3.4) points

compared to people with non-lung cancer, adjusted for the other factors in the model (Figure

2). Higher levels of fatigue and sleeping difficulty were also independently associated with

increased breathlessness when added to the final model, mean increase 0.24 (95% CI, 0.22 to

0.25) and 0.18 (95% CI, 0.16 to 0.21), respectively. They were not included in the final model

as they were thought to arise mainly as an effect, and not cause, of breathlessness. The

associations for functional status and concurrent symptoms were similar for differences

between patients and for changes within patients over time.

The adjusted change in breathlessness over time did not differ by age, gender, the presence of

malignancy, or lung involvement. Model estimates remained unchanged independent of

10

whether phase was included or removed from the model. Findings were similar when

restricting the analyses to patients with two or more breathlessness measurements.

Patients who increased two or more points of breathlessness distress in the final week were

more likely to be diagnosed with lung cancer, have higher AKPS, and tended to have higher

levels of difficulty sleeping, fatigue, and pain compared to other patients (Table 4).

On the day of dying, the level of breathlessness distress was highest in patients with

respiratory failure (mean 4.1 points; 95% CI, 3.0 to 5.3), followed by people with lung cancer,

cardiovascular disease, and other primary diagnoses (Figure 2).

DISCUSSION

Main findings

To our knowledge, this is the largest study to date of breathlessness in the final week of

life.(2, 9, 19) The main findings are that distress from breathlessness is frequently

encountered but that the population prevalence of 35% and average severity remains stable

over the last week of life. However, there is wide variation in the individual levels of

breathlessness. A number of factors were independently associated with higher

breathlessness: male gender; younger age; lung disease or a non-cancer diagnosis; concurrent

pain; higher levels of fatigue or nausea; higher rates of sleeping difficulty; better functional

status; and clinical instability. Lung involvement (by cancer or COPD) was associated with

the largest increase in breathlessness.

Strengths and limitations

11

Strengths of the present study include its prospective, national, multicenter design; large study

population; and collection of clinically relevant data on symptom distress using a validated

measurement tool (14, 20) at point of care in a real-world palliative care setting. The analysis

explored differences both between individuals and within individuals over time, and evaluated

a range of clinically relevant factors.

A limitation of the study is that PCOC collects clinical and symptom data only at each change

of phase. However, this anchors the measurements to changes in the patient’s clinical

condition. The phases have been validated and inter-rater reliability was good although

slightly lower for the deteriorating phase.(12) Changes that are clinically significant are likely

to trigger a change in phase and therefore to be recorded. The associations with phase in the

present study have good face validity, with lowest and highest levels of breathlessness in

stable and unstable phase, respectively. It is acknowledged that individual breathlessness

trajectories vary and might differ substantially from the group average.(2) An alternative

approach would have been to classify the individual trajectories using the change in scores

that is likely to be perceived as clinically important by the patient, such as 1 point on a 0-10

NRS of chronic breathlessness.(2, 21) That approach may be less feasible in the presence of

irregular or missing measurements and does not account for random measurement error or for

systematic differences in determinants of breathlessness between individuals or over time. We

used mixed effects linear regression which accounts for missing data and repeated

measurements.(18) Bias from data missing not at random was not supported as findings were

robust when analyzing only people with two or more measurements during the last week of

life. Although patients were asked to rate the severity of symptom distress,(14) ratings might

pertain to overall severity in some patients. Ratings of severity may however reflect mainly

symptom unpleasantness or distress in advanced disease, as reported in pain.(22) A possible

limitation is that we lacked data on whether ratings that were reported by a proxy. Patients

12

self-reported whenever possible but proxy ratings are inevitable as people deteriorate and

become unable to report at end of life. In this setting, proxy reports from carers and palliative

care professionals may both under- and over-estimate the presence and severity of

breathlessness but are useful.(15, 16) We also lacked data on which interventions and

treatments that the patients received. The findings should be interpreted in the context of

current clinical practice by specialist palliative care services at the end of life.

What this study adds

The present study extends previous observations in several important ways. Breathlessness

has been reported to increase as death approaches,(2, 9, 19) but those studies included almost

entirely populations with cancer with few data points in the last week of life. (2, 9) The

present larger dataset shows that the mean breathlessness distress remained unchanged at a

population level in the last week of life. This finding is a robust representation of

breathlessness in the last week of life and unlike some other studies on the topic (23, 24)

accounted for repeated measurements. The prevalence and severity of breathlessness were

consistent with a study during the last month of life in a similar palliative care setting.(9) The

lack of a change in intensity or prevalence of breathlessness over the last week of life in the

present study compared to previous reports would seem to be simply a matter of when in the

illness trajectory that breathlessness is worst.(19)

The present study confirms that patients with lung involvement experience markedly higher

levels of breathlessness distress than do people without lung involvement. (2, 9, 19, 25-31)

The present study identified a number of patient characteristics that are independently

associated with higher breathlessness distress in the last week of life. Patients with

nonmalignant disease and especially lung involvement suffered from markedly higher levels

13

of breathlessness. Airway obstruction, lung infection, pericardial and pleural effusions have

been associated with increased breathlessness near death,(10) likely mediated through

impaired gas diffusion in the lungs, increased ventilatory drive, breathing effort, and relative

ventilator insufficiency.(5) Despite being more symptomatic, patients with advanced

respiratory disease, and particularly end-stage COPD, may also receive less treatment with

opioids for breathlessness compared to cancer patients due to fear of respiratory

depression.(31) The findings of higher breathlessness in people with better functional status

may be related to higher mobility and hence more exertional breathlessness.

Interestingly, concurrent negative sensations – fatigue and nausea were independently

associated with increased breathlessness distress near death. This is consistent with previous

reports including studies of symptom clusters.(32-35) Fatigue and sleeping difficulty may

largely be consequences, and not causes, of increasing distress from breathlessness. The

presence of concurrent negative emotions have been associated with an increase in the

affective component of breathlessness − that breathlessness is perceived as more distressing

or frightening.(36, 37) This could be mediated through a co-activation of similar limbic

systems involved in the awareness of distressing breathlessness.(36) The question arises

whether relief of concurrent symptoms could indirectly decrease the severity and distress

from breathlessness. Optimized treatment of the underlying cause(s) is an important part of

breathlessness management, including treatment of symptomatic anemia, pleural effusions,

and airflow obstruction.(27, 38, 39) Relief of concurrent unpleasant symptoms such as nausea

and pain could be an important approach to relieve the distress from breathlessness.

Implications for research

14

A key question that arises from these findings is what happens to the people with the identical

diagnoses who are not seen by palliative care services. Are they not referred because their

level of symptom control is reasonable, or is it even more insufficient? Do symptom control

rates differ between the inpatient and the community setting?

It is not clear from this study what was being done to relieve breathlessness. Given that the

evidence base for the treatment of chronic refractory breathlessness is almost entirely derived

from people earlier than the last week of life, do the evidence for pharmacological and non-

pharmacological interventions apply to people near death? Understanding the actual

interventions to deal with breathlessness is important to ensure that the most effective

therapies are being used.

Implications for clinical practice

At a health systems level, it has been demonstrated that symptom control can be

systematically improved.(11) While rates of breathlessness may not increase in the last week

of life, breathlessness often remains insufficiently controlled. To improve the levels of

comfort in the last week of life, clinicians need to identify patients with the factors that are

independently associated with a higher risk of uncontrolled breathlessness in order to do two

things: focus greater efforts to ensure evidence-based symptom management interventions are

pro-actively implemented; and to discuss with the patient and his / her family the potential for

breathlessness to be present so that concerns they have can be addressed. Patients with end-

stage lung disease have particularly high frequency and severity of breathlessness. These

people deserve close clinical monitoring in order to optimize evidence-based treatments and

hence reduce distress at the end of life.(40)

15

DISCLOSURES

The authors have no conflicts of interest related to the present paper to disclose.

ACKNOWLEDGEMENTS

We thank Karen Quinsey, director of PCOC, for valuable feedback on the analysis.

16

References

1. Moens K, Higginson IJ, Harding R. Are there differences in the prevalence of palliative care-related problems in people living with advanced cancer and eight non-cancer conditions? A systematic review. J Pain Symptom Manage 2014;48:660-77. 2. Bausewein C, Booth S, Gysels M, et al. Individual breathlessness trajectories do not match summary trajectories in advanced cancer and chronic obstructive pulmonary disease: results from a longitudinal study. Palliat Med 2010;24:777-86. 3. Janssens T, De Peuter S, Stans L, et al. Dyspnea perception in COPD: association between anxiety, dyspnea-related fear, and dyspnea in a pulmonary rehabilitation program. Chest 2011;140:618-25. 4. Lansing RW, Gracely RH, Banzett RB. The multiple dimensions of dyspnea: review and hypotheses. Respir Physiol Neurobiol 2009;167:53-60. 5. Parshall MB, Schwartzstein RM, Adams L, et al. An official American Thoracic Society statement: update on the mechanisms, assessment, and management of dyspnea. Am J Respir Crit Care Med 2012;185:435-52. 6. Gysels MH, Higginson IJ. The lived experience of breathlessness and its implications for care: a qualitative comparison in cancer, COPD, heart failure and MND. BMC Palliat Care 2011;10:15. 7. Nauck F, Alt-Epping B. Crises in palliative care--a comprehensive approach. Lancet Oncol 2008;9:1086-91. 8. Currow DC, Ward AM, Plummer JL, Bruera E, Abernethy AP. Comfort in the last 2 weeks of life: relationship to accessing palliative care services. Support Care Cancer 2008;16:1255-63. 9. Currow DC, Smith J, Davidson PM, et al. Do the trajectories of dyspnea differ in prevalence and intensity by diagnosis at the end of life? A consecutive cohort study. J Pain Symptom Manage 2010;39:680-90. 10. Chiu TY, Hu WY, Lue BH, et al. Dyspnea and its correlates in taiwanese patients with terminal cancer. J Pain Symptom Manage 2004;28:123-32. 11. Currow DC, Allingham S, Yates P, et al. Improving national hospice/palliative care service symptom outcomes systematically through point-of-care data collection, structured feedback and benchmarking. Support Care Cancer 2015;23:307-15. 12. Masso M, Allingham SF, Banfield M, et al. Palliative Care Phase: inter-rater reliability and acceptability in a national study. Palliat Med 2015;29:22-30. 13. Abernethy A, Shelby-James T, Fazekas B, Woods D, Currow D. The Australia-modified Karnofsky Performance Status (AKPS) scale: a revised scale for contemporary palliative care clinical practice [ISRCTN81117481]. BMC Palliative Care 2005;4:7. 14. Aoun SM, Monterosso L, Kristjanson LJ, McConigley R. Measuring symptom distress in palliative care: psychometric properties of the Symptom Assessment Scale (SAS). J Palliat Med 2011;14:315-21. 15. Simon ST, Altfelder N, Alt-Epping B, et al. Is breathlessness what the professional says it is? Analysis of patient and professionals' assessments from a German nationwide register. Support Care Cancer 2014;22:1825-32. 16. Hui D, Morgado M, Vidal M, et al. Dyspnea in hospitalized advanced cancer patients: subjective and physiologic correlates. J Palliat Med 2013;16:274-80. 17. Oldenmenger WH, de Raaf PJ, de Klerk C, van der Rijt CCD. Cut Points on 0–10 Numeric Rating Scales for Symptoms Included in the Edmonton Symptom Assessment Scale in Cancer Patients: A Systematic Review. Journal of pain and symptom management 2013;45:1083-1093. 18. Graubard BI, Korn EL. Modelling the sampling design in the analysis of health surveys. Stat Methods Med Res 1996;5:263-81. 19. Seow H, Barbera L, Sutradhar R, et al. Trajectory of performance status and symptom scores for patients with cancer during the last six months of life. J Clin Oncol 2011;29:1151-8. 20. Richardson LA, Jones GW. A review of the reliability and validity of the Edmonton Symptom Assessment System. Curr Oncol 2009;16:55.

Formatted: English (U.S.)

17

21. Johnson MJ, Bland JM, Oxberry SG, Abernethy AP, Currow DC. Clinically important differences in the intensity of chronic refractory breathlessness. J Pain Symptom Manage 2013;46:957-63. 22. Clark W, Yang JC, Tsui S-L, Ng K-F, Bennett Clark S. Unidimensional pain rating scales: a multidimensional affect and pain survey (MAPS) analysis of what they really measure. Pain 2002;98:241-247. 23. Henoch I, Bergman B, Gustafsson M, Gaston-Johansson F, Danielson E. The impact of symptoms, coping capacity, and social support on quality of life experience over time in patients with lung cancer. J Pain Symptom Manage 2007;34:370-9. 24. Kroenke K, Johns SA, Theobald D, Wu J, Tu W. Somatic symptoms in cancer patients trajectory over 12 months and impact on functional status and disability. Support Care Cancer 2013;21:765-73. 25. Bruera E, Schmitz B, Pither J, Neumann CM, Hanson J. The frequency and correlates of dyspnea in patients with advanced cancer. J Pain Symptom Manage 2000;19:357-62. 26. Dudgeon DJ, Kristjanson L, Sloan JA, Lertzman M, Clement K. Dyspnea in cancer patients: prevalence and associated factors. J Pain Symptom Manage 2001;21:95-102. 27. Dudgeon DJ, Lertzman M. Dyspnea in the advanced cancer patient. J Pain Symptom Manage 1998;16:212-9. 28. Edmonds P, Higginson I, Altmann D, Sen-Gupta G, McDonnell M. Is the presence of dyspnea a risk factor for morbidity in cancer patients? J Pain Symptom Manage 2000;19:15-22. 29. Weingaertner V, Scheve C, Gerdes V, et al. Breathlessness, functional status, distress, and palliative care needs over time in patients with advanced chronic obstructive pulmonary disease or lung cancer: a cohort study. J Pain Symptom Manage 2014;48:569-81 e1. 30. Muers MF, Round CE. Palliation of symptoms in non-small cell lung cancer: a study by the Yorkshire Regional Cancer Organisation Thoracic Group. Thorax 1993;48:339-43. 31. Ahmadi Z, Lundström S, Janson C, et al. End-of-life care in oxygen-dependent COPD and cancer: a national population-based study. European Respiratory Journal 2015. 32. Clark N, Fan VS, Slatore CG, et al. Dyspnea and pain frequently co-occur among Medicare managed care recipients. Ann Am Thorac Soc 2014;11:890-7. 33. Tanaka K, Akechi T, Okuyama T, Nishiwaki Y, Uchitomi Y. Factors Correlated with Dyspnea in Advanced Lung Cancer Patients: Organic Causes and What Else? Journal of Pain and Symptom Management 2002;23:490-500. 34. Henoch I, Bergman B, Gustafsson M, Gaston-Johansson F, Danielson E. Dyspnea experience in patients with lung cancer in palliative care. Eur J Oncol Nurs 2008;12:86-96. 35. Fan G, Filipczak L, Chow E. Symptom clusters in cancer patients: a review of the literature. Curr Oncol 2007;14:173-9. 36. Laviolette L, Laveneziana P. Dyspnoea: a multidimensional and multidisciplinary approach. Eur Respir J 2014;43:1750-62. 37. von Leupoldt A, Mertz C, Kegat S, Burmester S, Dahme B. The impact of emotions on the sensory and affective dimension of perceived dyspnea. Psychophysiology 2006;43:382-6. 38. Ekstrom MP, Abernethy AP, Currow DC. The management of chronic breathlessness in patients with advanced and terminal illness. BMJ 2015;349:g7617. 39. Marciniuk DD, Goodridge D, Hernandez P, et al. Managing dyspnea in patients with advanced chronic obstructive pulmonary disease: a Canadian Thoracic Society clinical practice guideline. Can Respir J 2011;18:69-78. 40. Homsi J, Walsh D, Rivera N, et al. Symptom evaluation in palliative medicine: patient report vs systematic assessment. Support Care Cancer 2006;14:444-53.

Formatted: English (U.S.)

18

TABLES

Table 1. Characteristics of 12,778 patients in specialist palliative care

Characteristic All

n (%)

Age 73 (SD, 14)

Male gender 6,922 (54)

Malignant disease (N=12,720) 9,872 (78)

Primary diagnosis

Lung cancer 2,195 (17)

Colorectal cancer 1,168 (9)

Other gastrointestinal cancer 963 (8)

Breast cancer 715 (6)

Neurologic disease * 754 (6)

Hematological cancer 661 (5)

Pancreas cancer 614 (5)

Prostate cancer 552 (4)

Cardiovascular disease 554 (4)

Other non-malignancy 422 (3)

Respiratory failure 380 (3)

Other diagnoses † 3,742 (29)

State (no. services) (N=12,720)

New South Wales / ACT (22) 4,010 (31)

Queensland (21) 2,255 (18)

South Australia (12) 914 (7)

19

Tasmania (6) 587 (5)

Victoria (18) 3,219 (25)

West Australia (8) 1,793 (14)

Setting of care (N=12,776)

Hospital 9,885 (77)

Community 2,891 (23)

Phase

Stable 1,041 (8)

Unstable 1,503 (12)

Deteriorating 4,311 (34)

Terminal 5,923 (46)

AKPS, median (IQR) (N=12,111) 20 (20-40)

Data presented as n (%) unless otherwise specified. Baseline place of care and phase were for

the first record and AKPS was for all reports in the last week of life.

* Neurologic disease included stroke, dementia, and neuromuscular disease.

† Other diagnoses included end-stage kidney, liver, multiorgan failure, and other cancers.

Abbreviations: ACT = Australian Capital Territory; AKPS = Australia-modified Karnofsky

Performance Status; IQR = interquartile range; SD = standard deviation.

20

Table 2. Breathlessness measurements in 12,778 patients over the last week of life

Days before death Breathlessness measurements, n

7 1,610

6 1,808

5 2,153

4 2,548

3 3,094

2 3,936

1 5,145

0 1,763

Total 22,057

The total number of breathlessness measurements per day in the last week of life. Each patient

could have only one measurement per day. The number of measurements per patient during

the last week was a mean 1.7 (SD, 0.9; median 1; range 1 to 8).

21

Table 3. Multivariate model of breathlessness in 12,778 patients during the last week of

life

Variable Change in mean breathlessness 95% CI P-value

Day nearer death 0.14 0.12 to 0.17 <0.001

Male gender 0.12 0.02 to 0.22 0.022

Age (per y) -0.01 -0.01 to 0.00 0.001

Primary diagnosis

Cancer, non-lung Ref

Lung cancer 1.64 1.46 to 1.74 <0.001

Cardiovascular 1.40 1.15 to 1.64 <0.001

Respiratory failure 3.08 2.79 to 3.36 <0.001

Other non-cancer 0.35 0.20 to 0.49 <0.001

AKPS (per 10 points) 0.22 0.18 to 0.26 <0.001

Nausea (per point) 0.12 0.09 to 0.14 <0.001

Pain (per point) 0.17 0.16 to 0.19 <0.001

Phase

Stable Ref

Unstable 0.60 0.42 to 0.77 <0.001

Deteriorating 0.35 0.19 to 0.52 <0.001

Terminal -0.04 -0.23 to 0.15 0.690

Model intercept

Intercept 0.30 -0.08 to 0.68 0.117

Associations with the average breathlessness on a 0-10 numerical rating scale in palliative care

patients (n=9,759; 13,171 observations) during the last week of life. Associations are estimated using

multivariate random effects linear regression and are adjusted for all other variables in the model.

Abbreviations: AKPS = Australia-modified Karnofsky Performance Status; CI = confidence interval;

IQR = interquartile range; SD = standard deviation.

22

Table 4. Comparison of patients who increased ≥ 2 NRS points in breathlessness

anytime in the last week of life among 6,207 patients with multiple measurements

Characteristic Increased ≥ 2 NRS points

N = 2,886 (46%)

Other

N = 3,321 (54%)

Age 72.5 ± 13.5 73.3 ± 13.4

Male gender 1,617 (56) 1,782 (54)

Malignant disease 2,388 (83) 2,716 (82)

Primary diagnosis

Cancer, non-lung 1,726 (60) 2,240 (68)

Lung cancer 662 (23) 476 (14)

Cardiovascular 125 (4) 97 (3)

Respiratory failure 102 (4) 58 (2)

Other non-cancer 261 (9) 437 (13)

Hospital care 2,337 (81) 2,789 (84)

AKPS * 30 (20 to 40) 20 (20 to 35)

Difficulty sleeping * 0.5 (0 to 2.25) 0 (0 to 1.33)

Fatigue * 4 (2.50 to 5.67) 3 (0 to 5.50)

Nausea * 0 (0 to 1.00) 0 (0 to 0.67)

Pain * 2 (0.75 to 3.67) 1.67 (0 to 3.33)

Data presented as mean ± standard deviation or count (%) unless otherwise stated. Patients in

the increasing breathlessness group had slightly more lung cancer, lower functional status

(AKPS), and tended to have higher levels of difficulties sleeping, fatigue, and pain.

* Median score (interquartile range) during follow-up.

Abbreviations: AKPS = Australia-modified Karnofsky Performance Status; NRS = 0-10

numerical rating scale of symptom distress.

23

FIGURE LEGENDS

Figure 1: Reported levels of distress from breathlessness in the week before death

(n=12,778).

Severity categories were defined as none (NRS 0), mild (NRS 1 to 3), moderate (NRS 4 to 7),

and severe (NRS 8 to 10).

Abbreviation: NRS = numerical rating scale.

Figure 2: Mean breathlessness distress by diagnosis group (n=12,778).

Mean scores on a 0-10 numerical rating scale by diagnosis group, adjusted for gender, age,

Australia-modified Karnofsky Performance Status, concurrent nausea, pain, and clinical phase

using random effects linear regression.