Factors associated with quality of life in patients receiving palliative radiotherapy for bone metastases: a secondary cross-sectional analysis of data from a prospective multicenter observational study

Abstract

Objective

To identify factors significantly associated with quality of life (QOL) and determine if these associations are strong enough to predict certain aspects of QOL without measuring them.

Methods

We conducted an exploratory secondary analysis of baseline data of 224 patients (enrolled between December 2020 and March 2021) from a previously published prospective observational study on radiotherapy for bone metastases at 26 centres. Using univariable linear regression, we assessed the association between patient/treatment factors and QOL scale scores as measured by the European Organization for Research and Treatment of Cancer (EORTC) QOL Questionnaire Core 15-Palliative (QLQ-C15-PAL) and the EORTC QOL Questionnaire Bone Metastases module (QLQ-BM22).

Results

Age and sex were not significantly associated with QOL. Worse performance status, higher pain scores, and opioid and single-fraction use were significantly associated with most QOL scales; these four factors were associated with worse global QOL, worse functioning status, and more severe symptoms. The coefficients of determination for most QOL scales were less than 0.2, indicating that most of the variability in QOL scores was not explained by any of the explanatory variables.

Conclusion

Performance status, pain intensity, and opioid and single-fraction use were significantly associated with most QOL scales. However, the associations were not strong enough to estimate QOL.

Advances in knowledge

To date, the association between treatment factors and QOL in patients with bone metastases has not been fully studied. We identified the factors that were significantly associated with QOL and found that these associations were not strong enough to predict QOL.

Introduction

Bone metastases pose a serious burden on patients with advanced cancers. Various aspects of quality of life (QOL) can be impaired in patients with bone metastases. ^1,2 Since the main objective of palliative treatment for bone metastases is to improve the patient’s QOL, it is fundamental for health professionals to assess QOL and understand which aspects of the patient’s QOL are impaired. Radiotherapy is an important treatment for bone metastases that can improve the QOL of patients. ^3,4 However, the palliating effects of radiotherapy usually take several weeks to be exerted. ⁵ Moreover, radiotherapy alone is usually not sufficient to improve all aspects of QOL. ⁴ Healthcare professionals should provide necessary care to patients while simultaneously delivering palliative radiotherapy. If at least some of the QOL domains can be estimated from patient or treatment characteristics, the effective delivery of care may be accomplished without patients having to report their QOL.

Previous studies have explored the factors associated with QOL in patients with bone metastases. ^1,6,7 While these studies have identified significant patient-related factors, treatment factors such as analgesics and radiotherapy fractionation have not been studied to date. Additionally, it remains unclear how strongly QOL is associated with factors that are significant correlates, and whether any such factors can be used to estimate QOL. Therefore, in the present study, we sought to identify patient and treatment factors that are significantly associated with QOL and determine whether these associations are strong enough to estimate certain aspects of QOL without direct measurement.

Methods and materials

Patients and study design

The original prospective observational study was conducted at 26 centres, as previously reported. ⁸ The inclusion criteria in the original study were presence of: (1) signed informed consent and (2) planned radiotherapy for bone metastases. Only the patients who were deemed unsuitable for participation by the treating physician were excluded. Between December 2020 and March 2021, 333 patients received radiotherapy for bone metastases at the participating centres, of whom 232 (70%) were enrolled in the study. Of the 101 excluded patients, 11 patients (11%) refused to participate, the attending physician determined that the performance status of 68 (67%) was not good enough for the 6 month follow-up, the performance status was good in 13 (13%) but they were not fit for the follow-up, and 9 patients (9%) needed to start treatment before receiving an explanation about this study. After excluding five ineligible patients, one patient who withdrew from the study, and two patients who did not receive the planned radiotherapy, a total of 224 patients were included in the original study, and all were included in this analysis (Table 1).

Table 1.

Patient and treatment characteristics (n = 224)

Characteristic	No.	%
Age, years
Median (Range)	70 (28–89)
Sex
Female	85	38
Male	139	62
ECOG performance status
0	52	23
1	86	38
2	50	22
3	28	13
4	8	4
NRS worst pain score for the index pain a
Median (Range)	5 (0–10)
Missing	1	0.4
Opioid analgesic use at baseline
No	124	55
Yes	100	45
Fractionation
Single fraction	57	25
Multiple fraction	165	74
Missing	2	0.9

ECOG, Eastern Cooperative Oncology Group; NRS, numeric rating scale.

^aPain caused by the tumour planned to be irradiated.

QOL data were collected at enrolment, 2 month follow-up, and 6 month follow-up. Only QOL data at enrolment were used for the present cross-sectional analysis. The original prospective observational study was approved by the institutional review boards of the participating centres, and written informed consent was obtained from all patients. The study was registered with the University hospital Medical Information Network Clinical Trials Registry (URL, https://center6.umin.ac.jp/cgi-open-bin/ctr_e/ctr_view.cgi?recptno=R000048504) with the identification number UMIN000042491.

Evaluation

The patients completed the European Organization for Research and Treatment of Cancer (EORTC) QOL Questionnaire Core 15-Palliative (QLQ-C15-PAL) ⁹ and the EORTC QOL Questionnaire Bone Metastases module (QLQ-BM22). ¹⁰ The QLQ-C15-PAL is an abbreviated version of the QLQ-C30 used in palliative care settings. The QLQ-BM22 is a bone metastases module used to supplement the QLQ-C30. The QLQ-C15-PAL includes one global health status/QOL scale, two functional scales (physical and emotional functioning), and seven symptom scales (fatigue, nausea and vomiting, pain, dyspnoea, insomnia, appetite loss, and constipation). The QLQ-BM22 includes two functional scales (functional interference and psychosocial aspects) and two symptom scales (painful sites and pain characteristics). All items of the QLQ-C15-PAL and QLQ-BM22 were rated on a scale of 1–4, except for the global health status/QOL item of the QLQ-C15-PAL, which was rated on a scale of 1–7. All raw scores were converted to scores ranging from 0 to l00 according to the EORTC guidelines. ^11,12 Higher scores on the functional scales and global health status/QOL scale indicated better QOL, while higher scores on the symptom scales indicated worse QOL. Scale scores were calculated based on the average score of the answered items for scales where at least half of the items were answered. ^11,12 The exceptions to this rule were the scales of physical functioning, emotional functioning, fatigue, and nausea and vomiting of the QLQ-C15-PAL, which were abbreviated forms of the original scales from the QLQ-C30. For these four shortened scales, only one missing value in one item requires the scale to be treated as missing. For all scales of the QLQ-C15-PAL and QLQ-BM22, where less than half of the items were answered, the scale score was treated as missing. ^11,12

Additional information collected at enrolment included age, sex, Eastern Cooperative Oncology Group (ECOG) performance status, pain intensity, and opioid use. The treating radiation oncologist identified the pain caused by the irradiated tumour and recorded it as index pain. Information regarding the tumour causing the most severe pain was recorded when a patient had more than one painful irradiated tumour. Patients rated the intensity of their worst pain (in terms of the index pain) over the previous 3 days using an 11-point (0–10) numeric rating scale (NRS).

Statistical analysis

Univariable linear regression was performed to investigate the association between patient/treatment factors (explanatory variables) and QOL scale scores (dependent variables; treated as continuous variables). The a priori selected explanatory variables were age, sex, ECOG performance status, NRS pain score, opioid use, and fractionation. These factors were included because these were basic characteristics of patients (age and sex), significant correlates of QOL in past studies (performance status and pain score), ^1,6,7 and key treatment factors (opioid use and fractionation). Age was dichotomised at the median. The NRS pain score was dichotomised according to past studies, where pain with a score ≥8 was classified as severe. ^13,14 Available case analysis was used to address missing data, and linear regression analysis was performed after excluding patients for whom at least one dependent or explanatory variable was missing. The coefficient of determination value indicates the proportion of variance in the scale score explained by the explanatory variable; a value of 0 indicates no predictive value, and a value of 1 indicates perfect prediction. All statistical tests were two-tailed, and p < 0.05 was considered significant. Multiplicity adjustments were not performed in this study, considering the explorative nature of the present analysis. The statistical analyses were performed using R v. 4.2.2.

Results

Patients

Patient and treatment characteristics are shown in Table 1. The primary sites of the irradiated tumours (n = 224) were the lungs (n = 80; 36%), breast (n = 33; 15%), hepatobiliary/pancreas (n = 20; 9%), kidney/ureter (n = 19; 8%), prostate (n = 15; 7%), colorectum (n = 15; 7%), and others (n = 42; 19%). The sites of irradiated bone metastases (n = 303) were the spine (n = 157; 52%), pelvis, (n = 65; 21%), femoral bone (n = 30; 10%), ribs (n = 18; 6%), pectoral girdle (n = 13; 4%), and other sites (n = 20; 7%). Of the irradiated 157 spine sites, 65 were lumber, 58 were thoracic, 27 were cervical, and 7 were sacral spine.

Of the 224 patients analysed, 216 (96%) completed the planned radiotherapy and 2 (1%) did not. For the remaining six patients (3%), data on treatment completion were missing. The median total radiation dose was 22 Gy (range, 3–50.4 Gy) and the median number of fractions was 5 (range, 1–28). Data on fractionation were missing for two patients (0.9%). One-quarter of patients received single-fraction therapy (Table 1).

QOL scores

Few QOL scores were missing in the QOL scales (Table 2). For painful sites scale of the QLQ-BM22, 10 (4%) out of 224 patients had a missing score, whereas for the other scales, there were fewer missing data.

Table 2.

QOL scores

QOL scale	No.	Mean	Median	Interquartile range
QLQ-C15-PAL
Global health status/QOL a	217	45.2	50.0	33.3–66.7
Physical functioning b	217	55.5	60.0	26.7–93.3
Emotional functioning b	217	67.5	66.7	50.0–83.3
Fatigue c	216	49.7	44.4	33.3–66.7
Nausea and vomiting c	218	9.0	0.0	0.0–16.7
Pain c	218	54.7	50.0	33.3–83.3
Dyspnea c	218	23.2	16.7	0.0–33.3
Insomniac	218	35.6	33.3	0.0–66.7
Appetite loss c	217	36.4	33.3	0.0–66.7
Constipation c	218	29.8	33.3	0.0–33.3
QLQ-BM22
Painful sites c	214	27.4	26.7	13.3–33.3
Pain characteristics c	217	43.3	44.4	22.2–55.6
Functional interference b	217	56.6	58.3	37.5–75.0
Psychosocial aspects b	218	65.7	66.7	55.6–77.8

QLQ-BM22, QOL Questionnaire Bone Metastases module; QLQ-C15-PAL, QOL Questionnaire Core 15-Palliative; QOL, quality of life.

On a scale of 0–100, higher scores for the global health status/QOL scale and functional scales indicate better QOL, while higher scores for the symptom scales indicate more severe symptom and worse QOL.

^aA global health status/QOL scale.

^bA functional scale.

^cA symptom scale.

Association between patient/treatment factors and QOL scales

Among the six explanatory variables studied (Tables 3 and 4), one (0.4%) and two (0.9%) patients had missing values for the NRS pain score and fractionation, respectively. For the other four variables (age, sex, ECOG performance status, and opioid use), there were no missing values. Age was not significantly associated with QOL scores on any scales except for nausea and vomiting, where older patients experienced slightly milder symptoms (Table 3). Sex was not significantly associated with QOL scores on any of the scales. Worse ECOG performance status and higher NRS pain scores were significantly associated with worse global QOL, worse functioning status, and more severe symptoms on most QOL scales (Table 3). Opioid use at enrolment and single-fraction radiotherapy were associated with worse global QOL, worse functioning status, and more severe symptoms on most QOL scales (Table 4). For the majority of QOL scales, the coefficient of determination was <0.2, indicating that most of the variability in QOL scores could not be explained by any of the studied explanatory variables.

Table 3.

Associations between patient factors and QOL scale scores

QOL scale	Age (≥70 vs <70)			Sex (male vs female)			ECOG performance status (≥2 vs <2)			NRS worst pain score for the index pain (≥8 vs <8)
	β a	R2 b	p- value	β a	R2 b	p- value	β a	R2 b	p- value	β a	R2 b	p- value
QLQ-C15-PAL
Global health status/QOL	2.4	0.00	0.47	2.6	0.00	0.45	14.6	0.08	<0.001	14.2	0.06	<0.001
Physical functioning	4.6	0.01	0.28	8.6	0.02	0.052	39.9	0.38	<0.001	21.9	0.09	<0.001
Emotional functioning	3.7	0.01	0.30	5.6	0.01	0.12	19.8	0.14	<0.001	17.6	0.09	<0.001
Fatigue	1.4	0.00	0.71	1.9	0.00	0.62	18.3	0.10	<0.001	13.2	0.05	0.002
Nausea and vomiting	6.5	0.02	0.026	1.3	0.00	0.66	7.8	0.03	0.009	8.6	0.03	0.009
Pain	0.7	0.00	0.85	3.0	0.00	0.46	26.7	0.19	<0.001	28.7	0.18	<0.001
Dyspnea	1.6	0.00	0.67	5.8	0.01	0.13	5.3	0.01	0.17	6.9	0.01	0.10
Insomnia	5.3	0.01	0.25	0.5	0.00	0.92	15.1	0.05	0.001	24.0	0.10	<0.001
Appetite loss	3.2	0.00	0.46	0.9	0.00	0.85	21.2	0.10	<0.001	19.8	0.08	<0.001
Constipation	5.9	0.01	0.16	0.8	0.00	0.85	12.8	0.04	0.003	4.0	0.00	0.39
QLQ-BM22
Painful sites	1.6	0.00	0.49	1.4	0.00	0.58	8.1	0.05	<0.001	10.3	0.07	<0.001
Pain characteristics	4.2	0.01	0.25	2.1	0.00	0.57	12.3	0.05	<0.001	21.0	0.12	<0.001
Functional interference	2.2	0.00	0.55	0.9	0.00	0.81	16.6	0.09	<0.001	24.8	0.17	<0.001
Psychosocial aspects	0.6	0.00	0.80	4.3	0.01	0.10	14.0	0.13	<0.001	10.7	0.06	<0.001

ECOG, Eastern Cooperative Oncology Group; NRS, numeric rating scale; QLQ-BM22, QOL Questionnaire Bone Metastases module; QLQ-C15-PAL, QOL Questionnaire Core 15-Palliative; QOL, quality of life.

^aRegression coefficient (e.g. patients with pain score ≥8, on average, had 14.2 points lower scores in the global health status/QOL scale than those with pain score <8).

^bCoefficient of determination, which indicates the proportion of variance in the scale score explained by the respective factor (0 indicates no predictive value, whereas 1 indicates perfect prediction).

Table 4.

Associations between treatment factors and QOL scale scores

QOL scale	Opioid use (yes vs no)			Fractionation (single fraction vs multiple fraction)
	β a	R2 b	p-value	β a	R2 b	p-value
QLQ-C15-PAL
Global health status/QOL	10.6	0.05	0.002	1.3	0.00	0.73
Physical functioning	18.8	0.09	<0.001	9.6	0.02	0.049
Emotional functioning	9.6	0.03	0.006	9.0	0.02	0.025
Fatigue	12.0	0.05	0.001	14.1	0.05	<0.001
Nausea and vomiting	5.9	0.02	0.043	10.4	0.05	0.002
Pain	24.4	0.17	<0.001	14.1	0.04	0.002
Dyspnea	9.3	0.03	0.013	1.5	0.00	0.73
Insomnia	25.1	0.14	<0.001	16.7	0.05	0.001
Appetite loss	15.8	0.06	<0.001	15.5	0.04	0.002
Constipation	8.8	0.02	0.038	5.4	0.01	0.26
QLQ-BM22
Painful sites	8.6	0.06	<0.001	11.2	0.08	<0.001
Pain characteristics	19.5	0.13	<0.001	12.4	0.04	0.003
Functional interference	20.4	0.14	<0.001	10.5	0.03	0.012
Psychosocial aspects	8.1	0.05	0.001	7.0	0.03	0.016

EORTC, European Organization for the Research and Treatment of Cancer; QLQ-BM22, QOL Questionnaire Bone Metastases module; QLQ-C15-PAL, QOL Questionnaire Core 15-Palliative; QOL, quality of life.

^aRegression coefficient (e.g. patients who used opioids, on average, had 10.6 points lower scores in the global health status/QOL scale than those who did not use opioids).

^bCoefficient of determination, which indicates the proportion of variance in the scale score explained by the respective factor (0 indicates no predictive value, whereas 1 indicates perfect prediction).

Discussion

Worse performance status, intense pain, opioid use, and single-fraction use were significantly negatively associated with QOL on most QOL scales. However, none of these four variables sufficiently explained the variability of the QOL scores. Therefore, it may be difficult to estimate QOL based on these factors, and it is necessary to measure QOL to determine how the patient’s QOL is impaired. In the following paragraphs, we discuss the four factors that were shown to be significantly associated with QOL.

First, a worse performance status was significantly associated with worse QOL. This is in line with past studies that showed that a worse performance status is associated with worse global health status, ⁶ worse functioning, ^1,6,7 and more severe symptoms ^1,6,7 in patients with bone metastases. Since performance status is determined based on a patient’s daily living abilities, it is not surprising that it is associated with many QOL scales. Patients with impaired daily activities may naturally have a worse global QOL, worse functioning, and more severe symptoms.

Second, more intense pain was significantly associated with worse QOL on most scales. This is again in line with past studies that showed the association between more intense pain and worse global health status, worse functioning, and more severe symptoms. ¹ Another study investigating the association between pain intensity and the EuroQol 5-Dimension (EQ-5D) found correlations between pain intensity and the EQ-5D scales. ¹⁵ Therefore, pain intensity may be an important determinant of QOL in patients with bone metastases.

Our third finding was that opioid use was associated with a worse QOL. Considering that opioid use was associated with symptom scales regarding pain (pain [QLQ-C15-PAL], painful sites [QLQ-BM22], and pain characteristics [QLQ-BM22]), physicians may have prescribed opioid analgesics to patients experiencing more intense pain. As pain intensity is a determinant of QOL (as discussed in the previous paragraph), opioid use may be correlated with worse QOL. Moreover, the side-effects of opioid analgesics may have negatively influenced QOL scores. ¹⁶

The fourth factor identified as being related to QOL was single-fraction radiotherapy, which was associated with worse QOL. Since the determination of fractionation and QOL measurements were both performed before radiotherapy initiation, this association was not due to the effect of different fractionations, but to the physician’s choice of fractionation. Patients with poor functioning and severe symptoms may experience greater burden during daily treatment, and radiation oncologists, patients, and care givers may have negative attitudes towards multiple-fraction therapy. Prognostication may be another reason. Some radiation oncologists may choose multiple-fraction therapy to achieve prolonged pain response. ¹⁷ Patients with poor functioning and poor QOL tend to have a high risk of mortality and, thus, may tend to receive single-fraction therapy.

However, it is important to note that a significant correlation with QOL does not necessarily indicate that a given factor is useful for estimating QOL. Most previous studies on QOL in patients with bone metastases did not fully consider the strength of association. ^1,6,7 As demonstrated by the analysis of coefficients of determination in this study, none of the four factors identified as significant correlates had a sufficient predictive value for QOL.

The present study has several limitations. First, we did not enrol all the patients who received radiotherapy for bone metastases during the enrolment period. As patients with a better QOL tended to be enrolled in the study, this bias may have influenced the study results. However, a larger proportion (70%) of potentially eligible patients were enrolled in our study than in other studies, which reported enrolment proportions of 38.2% ¹ and 24%. ¹⁸ Second, although the proportions of missing data were ≤4% for all QOL scales, not all QOL scores were available. Last, multiplicity adjustment was not performed in this explorative study. Our findings should be confirmed in future studies.

In summary, performance status, pain intensity, and opioid and single-fraction use were significantly associated with most QOL scales. To our knowledge, the effect of treatment factors such as opioid use and fractionation on the QOL of patients receiving radiotherapy for bone metastases has not been studied previously. Our analyses of these factors may shed light on the interrelation between QOL and the treatment choice of physicians, patients, and care givers. However, the associations were not strong enough to predict QOL. To provide effective palliative care, it is important to measure QOL and identify specific aspects of QOL impairment.