Quality of life after the initial treatments of non-small cell lung cancer: a persistent predictor for patients’ survival

Lung cancer is considered the greatest contributor to death from cancer. It accounts for 1,180,000 deaths per year worldwide [1] and in France 26,624 deaths in 2005. In France, its incidence and mortality rates have decreased by 0.5% and 1.7%, respectively, since 1980 for men, but have increased for women [2]. Non-small-cell lung cancer (NSCLC) is accounted for 80% of these new cases [3, 4] and survival time for all stages of NSCLC (I, II, III-IV) is low with a median of 43.3 months [4]. Therefore, all prognostic factors must be identified in this group of patients to provide insights into the disease process and the therapeutic response.

Patient-reported outcomes, including health-related quality of life (HRQoL), symptoms, and functional status, are well established and useful in oncology for describing the clinical disease course, helping to select optimal treatment, or comparing populations of cancer patients with those having other diseases and with the general population [5‐9]. In general, HRQoL covers subjective perceptions of the positive and negative aspects of patient symptoms, including physical, emotional, social and cognitive functions, and importantly, disease symptoms and side effects of treatments [10]. Clinicians are increasingly considering HRQoL as critical to cancer patient care [11, 12].

Different instruments have been developed and validated [13, 14], and recently, researchers have begun to study the relation between HRQoL and patient’s survival. Often, patient HRQoL is measured after cancer diagnosis and before any treatment, and studies have shown different results. For example, measuring by the European Organization for Research and Treatment of Cancer (EORTC), Quality of Life Questionnaire Core-30 (QLQ C30), global HRQoL assessed before treatment, was found to be a strong predictor of survival in patients with NSCLC and lymph node abnormalities [15]. With this same instrument, another study showed that patient self-reported pain and dysphagia predicted overall survival in advanced NSCLC, independently of socio-demographic or clinical characteristics [16]. Fielding et al. used the Functional Assessment of Cancer Therapy-General instrument to measure patients with liver and lung cancer HRQoL in China and found that only physical well-being subscale predicted lung cancer survival [17]. However, Hernsdon et al. showed that after adjusting for clinical factors, the pain perceived by patients predicted survival, whereas overall HRQoL of the QLQ C-30 did not [18].

These heterogeneous results indicate that more research is needed to better understand the role of patient’s QoL on survival. In particular, the diagnostic process ending with delivery of the diagnosis and the treatment strategy that can be sources of psychological challenge [19, 20], and the early period of treatment that may add potential distress. Measuring HRQoL within months after the end of the diagnosis process reflects both the effect of the disease evolution and of the treatments during this period. A 3-month time frame period after the diagnosis process gives enough time to patients to undertake and receive some benefits from their treatments and at the same time to avoid a selection bias due to loosing too many patients. Whether the treatment received has levelled off the predictive value of HRQoL on survival remains unknown. We aimed to investigate whether HRQoL after the initial treatment still plays a role in predicting survival in patients with NSCLC. We used the Medical Outcomes Study Short Form 36 (SF-36) and the QLQ C-30 to evaluate HRQoL. The SF-36 has been useful in surveys of general and specific populations comparing the relative burden of diseases and differentiating the health benefits produced by a wide range of treatments [14, 21]. The QLQ C-30 allowed identifying health problems related to cancer and its treatments, such as pain, diarrhoea, nausea and vomiting, as well as appetite loss [13, 15].

This study showed that certain domains on health related quality of life, measured after the initial treatment, were related to the survival of patients with non-small-cell lung cancer. On the bivariate analysis, we observed that better physical- and emotional-role, social functioning and vitality as well as lesser pain symptom on SF-36 (score ≥50) were related to a better patients’ survival. On multivariate analysis, a persistent relation was observed for 2 domains: lesser (high scores) Bodily pain and better Social-functioning. The QLQ C-30 showed that on the bivariate analysis, better (high scores) physical-, role- and social functioning, and also global HRQoL as well as lesser symptoms (low scores) on fatigue, constipation, nausea-vomiting and appetite loss were related to a better survival. On the multivariate analysis, only a better Global QoL found to be related to a better patient’s survival. This relation was observed independently of known prognostic variables i.e. age, sex, stage of cancer, and initial treatments. Patients seemed to live few months longer with better perceived health.

This multicenter, prospective study involved patients with a wide variety of NSCLC severity (stage I to IV) while they were undergoing treatments. With a follow-up of 4 to 6 years, we could observe patient’s survival and whether HRQoL during treatments period continued to be related to survival. It is well known that both questionnaires that were used, the SF-36 and QLQ-C30, had robust psychometric properties. Also that HRQoL is one of measurements that revealed patients’ point of view regarding his/her health status and encompass the incidence and the extent of limitations due to physical capacity, which might occur because of the disease itself or the effects of treatment. Indeed, as stated by Franceshini et al., 54% of patients with lung cancer report dyspnoea, which contributes to the worsening of QoL, because this symptom might limit the ability to perform activities of daily living and to work [30].

Studies have shown different results on the relation between sex and survival of patients with NSCLC. The present study confirmed findings that women had a better survival [31‐34].

Readers need to be cautious in translating the results of this study to other populations. First, we measured HRQoL 3 months after the diagnostic process, so it represented patients’ perceptions of the overall early management of lung cancer, from diagnosis to therapy. However, this time period may introduce some selection bias because some patients had died or were lost to follow-up before HRQoL measurement. Patients who did not make it until a 3 month follow-up were older and mostly had a stage 4 of cancer. Among potentials responders, more than half (63.5%) returned back completed questionnaires, with few (from 1 to 5) missing data. Also, there was neither socio-demographical nor clinical characteristics difference between responders vs non responders. Second, previous studies showed poor performance status associated with worse survival and good performance with better survival [15, 16, 35]. We had information on treatments that were performed after the HRQoL measurement and we assumed, according to standards, that these treatments were prescribed in accordance with each patient’s condition, integrating performance status. Also, we did not dispose of information regarding the stage of treatment and therefore could not confirm the hypothesis that this variable may relate to survival.

Previous studies showed that better HRQoL scores measured at baseline (before any treatment) related to better survival of patients with cancer [15, 16, 36, 37]. This study confirmed previous ones that reported that better global QoL (QLQ C-30) as the strongest predictor of better survival [15, 36, 37]. However it did not find perceived pain as an independent predictor, as showed by another study for patients with stage III or IV NSCLC measured at diagnosis [16]. In the present study, we measured HRQoL 3 months after the diagnosis, when cancer management and treatments were ongoing. The mean scores on the SF-36 were increased by +0.5 to +6 points for patients who had some surgical treatment (n = 74) before HRQoL measurement than for those who did not. The results of our study suggested that, in addition to previous studies showing baseline HRQoL affecting survival, perceived lesser-pain and better social functioning measured by the SF-36 and perceived general QoL on QLQ C-30 at 3 months after the diagnosis may still affect survival. The better the HRQoL scores, the better the survival probability. As suggested by Gotay et.al., changes in HRQOL may be an early warning system that can be useful for clinical decision making, because HRQOL may deteriorate before disease progression is evident by other measures [36]. Although the observed impact of HRQoL may not be that strong, it plays a very important role besides other factors known that can influence these patients’ survival. However, further studies are needed to explore the fluctuation over time of HRQoL and whether this fluctuation, if it exists and depending on its time of measurement, is related to patient survival.

Measured after the initial treatment, this study showed a persistent relation between HRQoL and survival of patients with non-small lung cancer. Lesser perceived pain, a better perception on general health and general QoL are found to be related to a better chance to survive few months longer. This relation was observed independently of known prognostic variables i.e. age, sex, stage and initial treatments. Our results have implications for clinical practice. Clinicians may integrate these measures in their battery of indicators to continue monitoring patient health after the beginning of treatment and to carefully examine symptoms in self-rated HRQoL assessment in order to improve patients’ survival. If used in routine practice, it may also be useful to facilitate communication among clinicians, health professionals and patients to identify the most adaptable cancer management.