Prevalence of dyspnoea and usage of opioids in managing dyspnoea in advanced cancer patients: a longitudinal observational multi-centre study from India

Context Breathlessness is one of the devastating symptoms experienced by patients with advanced cancer and can be very challenging to manage. Objectives To find the point prevalence of dyspnoea in advanced cancer patients presenting to palliative care out-patient clinics, and the usage of opioids in palliation of dyspnoea. Methods We conducted a prospective observational study among all consecutive patients presenting to the outpatient clinics of six cancer centres in India from different parts of the country. In addition to routinely documented demographic and clinical data from patient charts, study investigators collected information on the Edmonton Symptom Assessment System, Cancer Dyspnoea Scale (CDS) and European Organization for Research and Treatment of Cancer Quality-of-Life Questionnaire Core 15 Palliative Care. We calculated the prevalence of dyspnoea and documented the usage of opioids in palliation of dyspnoea using tests of differences across patient characteristics. Results Between May 1, 2019, and April 30, 2020, 5,541 patients were screened for eligibility, and 288 were enrolled (48 patients from each of the six centres). We analysed the data of 288 patients, of which 36.4% had dyspnoea, with 28.5% with moderate to a severe degree (>4/10). Tiredness and loss of appetite were found to have associations with dyspnoea which were statistically significant on multivariate analysis. Standard palliative care management and routine usage of opioids preceded improvement in dyspnoea scores, CDS scores and quality of life scores throughout 7 days. Conclusion Dyspnoea is a common symptom in advanced cancer patients, presenting to outpatient clinics, and routine documentation of dyspnoea with appropriate usage of opioids helps in mitigation. Key message The article suggests that breathlessness is a common problem in advanced cancer patients and opioid prescription preceded symptom improvements in such patients.


Background and rationale
Dyspnoea or breathlessness is an uncomfortable and subjective feeling of breathing difficulty, which may vary in character and intensity for individual patients. This is emphasised by the consensus definition of the American Thoracic Society and it also suggests that this experience may be influenced and individualised by various physiological, psychosocial, behavioural and environmental factors and their interactions [1].
Dyspnoea is one of the commonly reported symptoms in patients with metastatic cancers and the prevalence is higher in patients with advanced malignancies, especially towards the end of life [2,3]. Progressive worsening of dyspnoea in a patient with advanced metastatic cancer is one of the markers of poor prognosis [4]. In a previous study in our centre, we found that the prevalence of dyspnoea in advanced cancers is as high as 44.37% [5]. The symptoms in cancer patients are often compounded with other symptoms and present in clusters. Some of the common symptoms that cluster with dyspnoea are fatigue, anxiety and depression, loss of overall wellbeing and functionality, leading to poor quality of life and higher caregiver burden [6,7]. This also has major implications in terms of further patient-care, disease management, treatment, outcomes and survival [8,9].
The clinical prevalence, quality and management of dyspnoea vary, and being a symptom, it is not measured often, leading to little data on this [10]. Dyspnoea can be highly subjective; thus, measurement requires asking the patient. In this study, participating palliative care units had existing processes to assess unidimensional assessments of dyspnoea for all outpatients as reported in a small sample from our clinic in a previous study [5]. In this study, we report the extent of the burden of dyspnoea in terms of its prevalence among a large sample of patients with advanced cancer in multiple regional centres across India.

Study population and methodology
We conducted a prospective observational multicentric study over 1 year between 2019 and 2020 in six high volume cancer centres from different regions of India. Tata Memorial Hospital (Mumbai) was the lead research site, and the others were Acharya Harihar Regional Cancer Centre Data were collected in the palliative medicine outpatient clinics by the study investigators. The average time for completion of the protocol was 15 minutes. Study participants were adults with advanced cancer who have been informed of their diagnosis and are well enough to complete the questionnaire. Any participants who were suffering from severe mental or cognitive disorders, and with potentially curable diseases referred for early palliative care were excluded.
Consecutive patients with advanced cancer referred to the outpatient clinics of the palliative care departments, and those who satisfy the eligibility criteria were considered for the study. Positive screened subjects were explained about the study and study procedure in a language they best understand. Participants were given adequate time to consider participation and decide. The Participant Information Sheet was given to them, and all questions were clarified to their satisfaction. Consent was taken by the Principal or Co-investigators of the participating institutions. In case of non-participation, the reason for non-participation was recorded and routine care was offered. At the participating centres, routine palliative care is provided by specialised palliative medicine multidisciplinary team of physicians, nurses, rehabilitation therapists and social workers who coordinate care that focuses on the patient and family's individual needs. The goal of such care is to provide the patient and their family with the best quality of life for as long as possible by assisting with the physical, emotional, social and spiritual stresses associated with life-limiting illness [11]. During the study, clinicians were able to start, adjust or stop other treatments for dyspnoea as per standard of practice according to clinical judgment. The study continued till the recruitment of a minimum sample size. The baseline data were recorded at presentation to the palliative care outpatient clinic. Those patients who were started on opioids as pharmacotherapy for breathlessness management were followed up on days 1, 3 and 7 to record the prescribing patterns and change in breathlessness scores.

Study variables
On the day of study enrolment (day 0), data were collected from a chart review of routinely collected variables like age, sex, place of residence, family income (per month), educational status, contact number, marital status, primary caregiver, primary cancer diagnosis, cancer stage (at assessment), date of diagnosis of cancer, sites of metastasis, comorbidities (if any), treatment received for primary cancer, Eastern Cooperative Oncology Group score [12], symptom assessment on the Edmonton Symptom Assessment System (ESAS) [13,14], current medications, and dosages. Two additional research questionnaires were used: Cancer Dyspnoea Scale (CDS) [15] and European Organization for Research and Treatment of Cancer Quality-of-Life Questionnaire Core 15 Palliative Care (EORTC QLQ-C15-PAL) [16] (details in Appendices). Follow-up, data were recorded on ESAS and CDS on day 1 and day 3 to record changes in symptom scores. On day 7, ESAS, CDS and EORTC QLQ-C15-PAL were recorded for change in symptoms and quality of life scores.

Statistical considerations
Calculation of minimal sample size was based on literature citing the improvement in dyspnoea scores in advanced cancer patients with morphine and study site-specific adjustments for data variability. The predicted standard deviation of the Visual Analogue Scale (VAS) from the literature is 16 mm (the VAS ranges from 0 to 100 mm) [17,18]. Thus, a minimum of 48 participants per centre could provide 80% power to detect a 10-mm difference in the scale, with α of 0.05, allowing for a 20% dropout rate [19].
We analysed data with descriptive statistics to find the point prevalence and regression methods with Network Visualisations [20] for the interrelationship of dyspnoea with other symptoms and quality of life. For modelling of the ESAS dyspnoea values over days 0, 1, 3 and 7, linear mixed-effects modelling was used to compare changes in scores between groups (with and without opioids) with adjustments for age, sex, study site and baseline dyspnoea level using CRAN lme4-package [21]. A first-order autoregressive structure was fit to adjust for autocorrelation among timepoints to account for individual variation in physiological or behavioural traits on repeated measures through time [22]. The model selection process used top-down strategy recommended by Zuur et al [23]. All analyses were performed on RStudio 2022.02.2 software [24]. All analyses used two-sided tests, and a two-sided p value of 0·05 or less was considered to be statistically significant.
The research was carried out in compliance with the Helsinki Declaration, and Ethics committee approvals were obtained from the Institutional Review Boards of all the participating centres and the trial was registered with the Clinical Trials Registry -India (CTRI/2021/10/037056).
Study participants consisted of 56.2% men, with 54.2% in the age range of 41-60 years. The most common cancers were genito-urinary in 17.9%, breast in 15%, head and neck in 13.4% and gastrointestinal in 13.1% of patients. 24.3% of patients had comorbidities, with 22% having a chronic pulmonary disease.
On ESAS, the most common symptoms were loss of well-being in 75%, tiredness in 74% and pain in 71% of patients. The quality of dyspnoea was recorded by CDS with a median value of 6.2 on the effort score, 3.8 on the anxiety score and 6.6 on the discomfort score. On the EORTC-QLQ-C-15-PAL scores for functional scales were 46.6, for symptom scales was 54.6 and for global health status was 54.4 (Table 1).
On multiple regression analysis between dyspnoea score and other items on ESAS, model validity is suggested by the p-value of the F-statistic is 1.312e-06, and significant predictor variables from the coefficients table are tiredness (t-statistic p-value = 0.04225) and loss of appetite (t-statistic p-value = 0.00475). The complexity of the interrelationships has been depicted in Figure 2 via network analysis of ESAS symptoms at baseline using force-directed plotting with Fruchterman-Reingold (normalised stress value = 0.18) ( Table 2).   Of the 288 patients with dyspnoea, 112 patients received oral morphine. Of these, 90 patients received opioids for pain, and no additional opioid was given for dyspnoea. Twenty-two patients received opioids for dyspnoea. Opioid consumption per day was recorded as Morphine Equivalent Daily Dose (MEDD) for patients on days 0, 1, 3 and 7. There was an increasing trend for MEDD over 7 days (Figure 3). A Pearson product-moment correlation was run to determine the relationship between opioid usage in dyspnoea and clinical demographic parameters. There was a strong, positive correlation with total score in CDS, which was statistically significant (r = 0.706, n = 22, p = 0.005).
The longitudinal results for change in opioid consumption, ESAS scores for dyspnoea, CDS scores and quality of life scores were calculated in 112 (38.9%) participants with moderate-severe dyspnoea, who received opioids and could complete all four observations on days 0, 1, 3 and 7. The mean dyspnoea scores on ESAS reduced from 6.1 on day 0 to 2.8 on day 7 post opioids and standard palliative care intervention ( Figure 4).
There was a reduction in effort, anxiety, discomfort and total scores of breathlessness on CDS over a follow-up period of 7 days ( Figure 5).
Quality of life scores was recorded on day 0 and day 7 to identify the effect of dyspnoea management on functional scores (daily life activities), symptom scores and global health scores (overall wellbeing) of the patient. The improvement was noticed in functional scores and global health status and there was a reduction in symptom scales on day 7 on EORTC-QLQ-C15-PAL scores ( Figure 6).

8
Overall, improvement in all the scores (ESAS dyspnoea scores, CDS scores, quality of life scores and MEDD) over 7 days showed a statistically significant result (p < 0.05). However, a mixed-model analysis showed that ESAS dyspnoea scores did not differ statistically between the groups (with and without opioids) by day 7 (Table 3).

Discussion
This is the first study determining the prevalence of breathlessness among palliative care outpatient clinic patients in major cancer centres in India from different parts of the country [25]. As many as 36.4% of patients attending the clinics present with breathlessness and onequarter of them report severe breathlessness. It was heavily related to other symptoms, and thus was one of the most important symptoms in study participants.
Large-scale measurement of dyspnoea in clinics is feasible. Across all the participating centres, study investigators suggested that subjective assessment of dyspnoea and correct documentation was an important part of the patient-centred care approach. They also noticed that the process was less time-consuming, and that the unidimensional dyspnoea assessment measures using the 0-10 Likert scale were efficient for screening dyspnoea in busy outpatient settings. However, multidimensional assessments are preferred in specialised palliative care practice, which is essential for a better understanding of the factors contributing to the patient's overall symptom experience [26]. The study results of this study were based upon the data obtained from the patients directly attending palliative medicine outpatient clinics over 1 year. In a study by Baker et al [27], the authors described similar findings and suggested that palliative care nurses can use unidimensional and multidimensional measures to assess and characterise breathlessness in busy clinical settings.
In this study, data collection and symptom assessment were done while the patients were seated for a while after arrival at the clinic. Hence, logically these ratings suggest dyspnoea in the non-exertional and resting condition of the patient. It was noticed that patients with cardiorespiratory system involvement due to cancer or other co-morbidities like chronic obstructive pulmonary disease, heart diseases and chronic kidney disease reported higher intensity of dyspnoea, with a score > 4/10. Also, those patients who reported pain score > 4/10 (n = 123), and fatigue score > 4/10 (n = 101) at the time of assessment had a 2.8-and 2.3-times higher prevalence of dyspnoea, respectively. These patients also reported higher intensity of dyspnoea. This depicts the impact of comorbidities and the presence of other symptoms on the patient's overall experience of dyspnoea.
ESAS estimates the level of dyspnoea to examine the average symptom intensity over the past 24 hours. This recalled dyspnoea rating is subjective, as it varies depending on the amount of physical exertion and overall experience during the interval time, and physiologically represents cardiorespiratory demand about delivery [28]. Most patients reported that their experience of the worst dyspnoea was not related to the level of exertion. This should not be surprising, as the study sample comprised of cases with advanced cancers with a conglomeration of other symptoms related to dyspnoea.
The reported prevalence of dyspnoea in advanced cancer patients is 36.4% in this study, which is comparable to a meta-analysis that included more than 10,000 patients with advanced cancer, with 10%-70% of patients reporting dyspnoea [2]. This difference can be due to the difference in the assessment of dyspnoea. Most studies in the literature assessing the prevalence of dyspnoea have used the Medical Research Council (MRC) dyspnoea scale [29]. The MRC identifies the level of breathlessness based on the level of exertion that provokes breathlessness and is a more objective measure. We used more subjective assessment tools, to determine the overall perception of dyspnoea.
The American Society of Clinical Oncology in their 2021 guideline for management of dyspnoea in advanced cancer recommended the usage of systemic opioids for patients who derive inadequate relief from nonpharmacologic interventions [30]. In a dose-finding pilot double-blind randomised clinical trial by Hui et al [31] prophylactic fentanyl improved breathlessness and walk distance (pre-post analysis) compared with placebo. Our study design unfortunately does not allow us to claim that opioid usage led to alleviation of dyspnoea using a mixed-model analysis; however, we documented that the usage of opioids preceded dyspnoea alleviation here. This is significant for Indian settings, as apart from the fear of respiratory depression and accelerated death with the usage of opioids [32], there are barriers to opioid usage/availability like bureaucratic hurdles and sociocultural/infrastructure challenges [33].
Our study data are restricted to palliative care outpatient clinics, and thus do not include many patients who attend oncology clinics. In high-risk groups with cardiac and pulmonary comorbidities, literature reports an increased prevalence of dyspnoea, up to 50%-70% [34]. We noticed differences in the prevalence of dyspnoea based on patient demographics like gender and age. Both patient and clinician gender influence the expectation of dyspnoea and its treatment [35]. These findings were consistent with the findings of Santos et al [36], who reported a higher prevalence of breathlessness among female gender and elderly patients.
We used clinic data for research purposes with necessary approval, which is considered real-life data and superior to administrative data. All patients provided a numerical rating for ESAS symptoms including dyspnoea. Such ordinal scoring captures full information, rather than a binary yes/no. For binary response choices, a participant may need some specific intensity of breathlessness to respond 'yes' to the presence of sensation and this magnitude may vary from person to person in different situations [37]. Our analyses cannot account for biases related to reporting, perceived experience or documentation.
ESAS data were collected by research staff including trained nurses and doctors. The data represents the real-world scenario and can be related to practical experiences in outpatient clinics: 1) Dyspnoea, being a subjective experience, should be patient-reported. However, doctors and nurses use some objective measures for respiratory distress to supplement patients' reporting of breathlessness. One of the reasons for this is perceived difficulty for patients to quantify the experience of breathlessness in a numeric scale or in cases where patient is unable to vocalise. Hence, it is worthwhile to use some objective measures of 'observed respiratory distress' in addition to the subjective reporting, to have more robust information.
For all the study participants who complete the reporting, we acknowledge that 10%-15% of participants may find it difficult to translate their perceived symptom experience to a numeric value. Ideally, such patients are 'poor raters' and should be excluded from the actual data analysis. We acknowledge the error in measurement that might have crept in due to their presence in the study population. The study investigators across all participating centres agreed with the fact that these 'poor raters' were the patients who are unable to numerically rate all the subjective experiences like pain, breathlessness and other symptoms. This is noteworthy while assessing the patients in routine clinical practice, to be aware of false reporting.
2) Ideally, dyspnoea assessment should be done at the first visit to the hospital, but such assessments are challenging and were out of scope of this study. The prevalence of dyspnoea as indicated in this study may not be the true representation of patients presenting to hospital with breathlessness. It might have been slightly under-estimated as some of the severely breathless patients, who are unstable at presentation, will be shifted to the emergency department upon arrival, to receive treatment before the study enrolment could be made. To help circumvent this coverage gap, both ESAS and CDS include an item for assessment of dyspnoea in past few days before presentation to hospital, recall for dyspnoea experience in this period is dependable.

Conclusion
Dyspnoea is one of the most distressing symptoms in patients with cancer. In this study, a third of the patients presented with dyspnoea of moderate to severe intensity. Cancer-associated dyspnoea includes multiple causes ranging from cancer itself to multiple ongoing comorbidities, respiratory and cardiovascular causes being the major ones. Although dyspnoea is less prevalent than pain, it should be routinely assessed as it significantly adds to both physical and emotional suffering. It is evident that assessment and documentation of dyspnoea are easy, less time-consuming and can be accurately done by all healthcare professionals, including doctors and nurses and its management significantly impacts overall patient experiences. The usage of opioids preceded dyspnoea alleviation in this study. We recommend that dyspnoea assessment should be routinely included as a part of the patient care plan, and opioids should be opted for palliation of dyspnoea.

Disclosures and conflicts of interest
The authors have no conflicts of interest to disclose. De-identified individual participant data underlying the results reported in this paper will be available, for 5 years after publication, to investigators for uses approved by the institutional review board of the Tata Memorial Hospital. Proposals should be directed to the corresponding author. Acceptance of a data access agreement will be required for access.

Author contributions statement
AG, AD, MAM and JD contributed to the conception or design of the work; AG, AD, PS, SM, VV, AGr, PSa and SB helped in data collection; AG and AD contributed to data analysis, interpretation and drafting the article; AG, AD, MAM, PS, JD, SM, VV, AGr, PSa, SB helped in critical revision and drafting the final version of the article to be published.