Baseline Quality of Life is Associated with Survival Among People with Advanced Lung Cancer

Abstract

Introduction

Lung cancer patients presenting with advanced cancer face low survival rates and a high symptom burden. There have been mixed findings for the association between survival and various patient reported outcomes (PROs).

Methods

We used prospective data from 111 lung cancer patients with stage III/IV disease to investigate the association of survival with PROs (European Organization for Research and Treatment of Cancer Core-30 and Lung Module). Cox proportional hazard models were used to examine the individual association between several PRO measures and survival.

Results

Pain in chest and global quality of life (QoL) were found to have the strongest association with survival with a 20% increased hazard of death per 10% increase in pain in chest and 14% decrease in hazard of death per 10% increase in global QoL.

Conclusion

Our results provide more evidence for the value of PRO data to inform clinical and patient decision-making.

Introduction

Lung cancer is the leading cause of cancer death in the world¹. The 5-year survival for lung cancer patients in countries with a high development has been historically dismal with estimates of 4–7% for those presenting with advanced stage cancer^1,2. Quality of life (QoL) has been extensively studied as a prognostic factor of survival in cancer patients^3–8. It is a particularly important consideration among patients with advanced stage lung cancer, where survival is expected to be the shortest. While global QoL has been found to be a good predictor of survival independent of receiving standard of care⁵ some studies have found conflicting results for particular items within the functional and symptom domains of QoL, in terms of their association with survival^6,8. Social functioning, role functioning, physical functioning, dyspnea, pain, and dysphagia have all had mixed results^4,6,8.

We investigated the association of survival with patient reported outcomes (PROs), including QoL and items within the functional and symptom domains as measured by the European Organization for Research and Treatment of Cancer Core (EORTC-QLQ-C30) and the Lung module (LC-13 subscale)^9,10. Both the EORTC-QLQ-C30 and the LC-13 subscale are validated questionnaires among cancer patients and the LC-13 scale is a lung cancer-specific module.⁹ We used baseline data from a randomized controlled trial evaluating whether a structured physical activity intervention can reduce fatigue in ambulatory patients with advanced lung cancer compared to usual care¹⁰.

Materials and Methods

Study Population/Patient Characteristics

Details regarding patient eligibility and treatment have been previously described¹⁰. Briefly, the study population (n=111) consisted of patients with a histological diagnosis of non-small cell lung cancer or small cell lung cancer, with incurable stage III/IV disease. Overall, 45% of the participants were female, with a median age of 64 years (range 34 to 80), median time from diagnosis of 8 months and a median survival time of 60.5 weeks (95% CI: 57.7, 83.1)¹⁰.

Variables and statistical analysis

QoL, functional, and symptom domains were measured by the European Organization for Research and Treatment of Cancer Core (EORTC-QLQ-C30) and the Lung module (LC-13 subscale)⁹. Unadjusted and adjusted (sex, age and randomization arm were selected a priori) Cox proportional hazard models were used to examine the association between each of the QoL measures individually and survival. The number of covariates in adjusted models were constrained by the number of deaths that occurred during follow-up. PRO measures included global QoL (2 items); functional scales (physical [5 items], role [2], emotional [4], cognitive [2], and social [2]); symptom scales (fatigue [3], pain [2], dyspnea [1], appetite loss [1], nausea/vomiting [2], insomnia [1], constipation [1], diarrhea [1], and financial problems [1]); and LC-13 symptom scales (coughing [1], dysphagia [1], dyspnea [3], hemoptysis [1], alopecia [1], pain in arm [1], pain in chest [1], peripheral neuropathy [1], pain in other parts [1], and sore mouth[1]), not including the item on pain medication which is not scored. Higher values in the symptom scale indicate worse symptoms, higher values in functional scales and global QoL indicate better function and QoL. We estimated the hazard ratio per 10% increase for each PRO measure. Time was calculated as time from randomization to death. Patients alive at the end of the study period were censored. The c-index was used to measure goodness of fit in adjusted models. Values can range from 0.5–1 with 1 being the best possible fit¹¹. All analyses were performed in SAS 9.4.

Results

The unadjusted and adjusted Cox proportional hazards models included 109 patients for most PROs, and 104 patients for the “pain in other parts“ item. Global QoL, physical functioning, role functioning, dyspnea (in both symptom scale and LC-13 symptom scale), appetite loss, pain in arm or shoulder, and pain in chest had statistically significant association (Table 1). Unadjusted and adjusted estimates were similar. Following adjustment for sex, age and randomization arm, physical functioning was no longer statistically significant.

Table 1:

Baseline Characteristics for Patients with Advanced Lung Cancer in the Impact of Physical Activity on Fatigue and Quality of Life Randomized Controlled Trial.

Baseline Characteristic	N (%)
Sex
Male	61 (55%)
Female	50 (45%)
Age (years) Median (range)	64 (34 – 80)
Months since diagnosis
<2	9 (8%)
2.1–4	19 (17%)
4.1–8	26 (24%)
8.1–16	28 (25%)
>16.1	29 (26%)
ECOG performance status
0	61 (55%)
1	46 (41%)
2	4 (4%)
Current anti-cancer treatment:
Not on active treatment	24 (22%)
On chemotherapy only	50 (45%)
Single agent/doublet	15/29
Chemotherapy regimen unknown	6
Combination chemotherapy+targeted therapy	6 (5%)
Targeted therapy along	31 (28%)
Chest Radiotherapy	0
Smoking history
Never smoker	42 (38%)
Current or ex-smoker	69 (62%)
Non-small cell lung cancer	106 (96%)
Small cell lung cancer	5 (4%)

Pain in chest was found to be the measure that had the strongest association with survival, with a 10% increase in pain in chest associated with a 20% increased hazard of death. A 10% increase in appetite loss was associated with an 8% increase in hazard of death. A 10% increase in symptom scale dyspnea and LC-13 dyspnea were associated with a 13% and 16% increase in hazard of death respectively. A 10% increase in pain in arm or shoulder had a 9% increase in hazard of death (Table 2). A 10% improvement in global QoL and role functioning were associated with a 14% and 11% decrease in hazard of death, respectively. The c-indices for the adjusted models ranged from 0.54 and 0.63.

Table 2:

Results of 22 unadjusted and 22 adjusted* models evaluating the association of each individual baseline QoL measures with survival. Hazard ratios represent a 10% increase in QoL measure.

Variable	Number of items	Unadjusted HR (95% CI)	P-value	Adjusted* HR (95% CI)	P-value	C-Index
Global QoL	2	0.88 (0.79, 0.97)	0.0098	0.86 (0.78, 0.95)	0.0038	0.60
Functional Scales
Physical Functioning	5	0.90 (0.81, 0.99)	0.045	0.90 (0.81, 1.01)	0.068	0.58
Role Functioning	2	0.89 (0.82, 0.97)	0.0057	0.89 (0.81, 0.96)	0.0037	0.60
Emotional Functioning	4	0.91 (0.83, 1.01)	0.088	0.92 (0.83, 1.01)	0.088	0.56
Cognitive Functioning	2	0.93 (0.84, 1.03)	0.16	0.93 (0.84, 1.03)	0.14	0.54
Social Functioning	2	0.92 (0.84, 1.00)	0.059	0.91 (0.84, 1.00)	0.046	0.58
Symptom Scales
Fatigue	3	1.06 (0.97, 1.16)	0.20	1.06 (0.97, 1.16)	0.19	0.58
Pain	2	1.08 (0.99, 1.17)	0.057	1.08 (0.99, 1.17)	0.078	0.59
Dyspnea	1	1.13 (1.04, 1.21)	0.0022	1.13 (1.05, 1.23)	0.0015	0.60
Appetite Loss	1	1.08 (1.01, 1.16)	0.024	1.07 (1.00, 1.15)	0.044	0.60
Nausea/Vomiting	2	1.03 (0.89, 1.19)	0.71	1.02 (0.88, 1.18)	0.84	0.55
Insomnia	1	0.99 (0.92, 1.06)	0.72	0.99 (0.93, 1.07)	0.88	0.54
Constipation	1	1.09 (1.01, 1.18)	0.037	1.09 (1.01, 1.18)	0.037	0.56
Diarrhea	1	0.93 (0.83, 1.03)	0.17	0.94 (0.84, 1.05)	0.28	0.56
Financial Problems	1	1.01 (0.94, 1.09)	0.77	1.03 (0.96, 1.11)	0.43	0.56
LC-13 symptom scales
Coughing	1	1.08 (0.99, 1.17)	0.055	1.09 (1.00, 1.18)	0.051	0.57
Dysphagia	1	1.07 (0.96, 1.20)	0.20	1.08 (0.97, 1.22)	0.16	0.58
Dyspnea	3	1.15 (1.03, 1.28)	0.012	1.15 (1.04, 1.28)	0.0078	0.61
Hemoptysis	1	1.17 (0.99, 1.39)	0.071	1.18 (1.00, 1.40)	0.057	0.56
Alopecia	1	1.07 (0.99, 1.16)	0.11	1.07 (0.99, 1.17)	0.10	0.56
Pain in Arm or Shoulder	1	1.09 (1.01, 1.18)	0.029	1.09 (1.01, 1.18)	0.034	0.59
Pain in Chest	1	1.16 (1.05, 1.29)	0.0035	1.20 (1.08, 1.33)	0.0006	0.63
Peripheral Neuropathy	1	1.09 (0.99, 1.18)	0.066	1.08 (0.98, 1.18)	0.13	0.56
Pain in Other Parts	1	1.01 (0.94, 1.09)	0.76	1.00 (0.93, 1.08)	0.99	0.56
Sore Mouth	1	0.92 (0.82, 1.04)	0.17	0.90 (0.80, 1.02)	0.10	0.59

^*Adjusted for age, sex and randomization arm.

Discussion

We performed secondary analysis of trial data to investigate the association between patient reported QoL, functioning and symptoms and survival in persons with advanced lung cancer prior to randomization in a physical activity study. We found that increased baseline global QoL and role functioning were associated with decreased hazard of death. Dyspnea, anorexia and particularly chest pain, often increase with progression of disease pain. We found increased pain in chest, dyspnea, pain in arm or shoulder and appetite loss were associated with increased hazard of death. Pain in chest had strongest association with a 20% increase in hazard of death per 10% increase in the LC-13 symptom scale.

Although almost all of our study population had metastatic non-small cell lung cancer they were not typical of the advanced lung cancer population, with 96% having an Eastern Co-operative Oncology Group (ECOG) performance status of 0 – 1¹⁰. Similarly, participants’ physical function and symptoms were better than expected for patients with advanced lung cancer with less deterioration over time. Median time from diagnosis to randomization was 8 months, at a time when the median survival for advanced lung cancer was 9–12 months, suggesting a selection bias resulting in a healthier population than typical in newly diagnosed lung cancer patients. This could impact the generalizability of the results.

Our findings are similar to a previous study by Efficace et al⁴ in patients with advanced non-small cell lung cancer evaluated prior to commencing chemotherapy treatment on clinical trials. Efficace et al also showed that global QoL, dyspnea, and appetite loss were associated with survival, however, they found physical functioning, social functioning, pain, and dysphagia were not. These differences may be, in part, due to differences in sample demographics with our study population having a higher proportion of females (45% vs 35%) and a median age that was 7 years older than in the Efficace study⁴. However, both studies had patients of good performance status suggesting that participants in clinical trials are often not representative of the advanced lung cancer population¹².

Despite limitations in population generalizability and the number of variables we were able to adjust for due to the number of events in our population, the findings reported here provide more evidence for the importance of collecting PRO data to inform clinical decision making. Our results provide further evidence for the association between PRO measures and survival.

Data availability

Data generated and analyzed for the current study are not publicly available but are available from the corresponding author upon reasonable request.