Advertisement

Anxiety After Diagnosis Predicts Lung Cancer–Specific and Overall Survival in Patients With Stage III Non–Small Cell Lung Cancer: A Population-Based Cohort Study

Open ArchivePublished:January 04, 2017DOI:https://doi.org/10.1016/j.jpainsymman.2016.12.338

      Abstract

      Context

      The question as to whether anxiety and depression are related to mortality in patients with lung cancer is inconclusive.

      Objectives

      Therefore, the present study is examining associations of anxiety and depression in a large representative sample of patients with Stage III non-small cell lung cancer.

      Methods

      Patients (n = 684) were routinely assessed for anxiety and depression with the PsychoSocial Screen for Cancer questionnaire after diagnosis of lung cancer and before treatment initiation between 2004 and 2010. Survival data were retrieved in May 2012. Cox proportional hazards regression analyses had been used as statistical procedures allowing adjustment for demographic, biomedical, and treatment variables.

      Results

      In analyses controlling for demographic, biomedical, and treatment prognosticators, anxiety but not depression was associated with increased lung cancer–specific (hazard ratio 1.04; 95% confidence interval 1.01–1.07; P = 0.035) and all-cause (hazard ratio 1.04; 95% confidence interval 1.01–1.07; P = 0.005) mortality. Secondary analyses revealed a confounder effect of performance status on the association between depression and mortality, such that the removal of performance status identified a significant relationship of depression on lung cancer–specific and all-cause mortality.

      Conclusion

      In a large population-based sample of patients with non–small cell lung cancer analyses demonstrated associations of anxiety with mortality, adding to the evidence that psychosocial factors might play a role in disease progression in this patient group. Because emotional distress is associated with continued smoking and lack of success of smoking cessation attempts, psychological interventions potentially could influence length of survival in lung cancer patients.

      Key Words

      Although the incidence of lung cancer is decreasing,
      • Govindan R.
      • Page N.
      • Morgensztern D.
      • et al.
      Changing epidemiology of small-cell lung cancer in the United States over the last 30 years: analysis of the surveillance, epidemiologic, and end results database.
      lung cancer remains the third most frequent type of cancer.
      • Siegel R.L.
      • Miller K.D.
      • Jemal A.
      Cancer statistics, 2015.
      Most patients with lung cancer are former or current smokers,
      • Alberg A.J.
      • Brock M.V.
      • Ford J.G.
      • Samet J.M.
      • Spivack S.D.
      Epidemiology of lung cancer: diagnosis and management of lung cancer, 3rd ed.: American College of Chest Physicians evidence-based clinical practice guidelines.
      and second-hand smoke and other environmental factors such as air pollution and radon exposure are also risk factors.
      • de Groot P.
      • Munden R.F.
      Lung cancer epidemiology, risk factors, and prevention.
      Non–small cell lung cancer is associated with a poor one-year net survival ranging from only 30% to 46% in various western countries.
      • Walters S.
      • Maringe C.
      • Coleman M.P.
      • et al.
      Lung cancer survival and stage at diagnosis in Australia, Canada, Denmark, Norway, Sweden and the UK: a population-based study, 2004-2007.
      Not surprisingly, patients with lung cancer are considerably emotionally distressed. In a large epidemiologic study, lung cancer patients reported the greatest levels of anxiety and depression after diagnosis
      • Linden W.
      • Vodermaier A.
      • Mackenzie R.
      • Greig D.
      Anxiety and depression after cancer diagnosis: prevalence rates by cancer type, gender, and age.
      with patients who were metastasized reporting the greatest distress.
      • Vodermaier A.
      • Linden W.
      • MacKenzie R.
      • Greig D.
      • Marshall C.
      Disease stage predicts post-diagnosis anxiety and depression only in some types of cancer.
      Feelings of guilt and stigma as a result of having provoked the disease with one's smoking behavior may contribute to distress on top of the poor prognosis.
      • Chambers S.K.
      • Dunn J.
      • Occhipinti S.
      • et al.
      A systematic review of the impact of stigma and nihilism on lung cancer outcomes.
      Although a diagnosis of cancer is predictably perceived as a major health threat, subjective patient responses to the same diagnosis vary.
      Psychoneuroimmunologic models assume a mechanism by which anxiety and depression can affect survival in cancer patients. Under this framework, psychological processes such as anxiety and depression follow a pathway ultimately resulting in impaired nervous system and immune function, which in turn can promote preterm death.
      • Ader R.
      • Cohen N.
      • Felten D.
      Psychoneuroimmunology: interactions between the nervous system and the immune system.
      • Reiche E.M.
      • Nunes S.O.
      • Morimoto H.K.
      Stress, depression, the immune system, and cancer.
      The question whether emotional distress affects survival in cancer patients is of scientific interest since decades and had been studies in mixed samples of cancer patients and distinct cancer types. In patients with lung cancer, earlier studies showed conflicting findings, possibly because of methodologic limitations such as small sample size, poor sample aggregation (e.g., some studies combined small cell and non–small cell lung cancer without adequate control), and lack of biomedical and treatment descriptors and their control. In a small sample of 40 Stage IV lung cancer patients, an association existed between initial quality of life and survival,
      • Ganz P.A.
      • Lee J.J.
      • Siau J.
      Quality of life assessment. An independent prognostic variable for survival in lung cancer.
      whereas another
      • Herndon J.E.
      • Fleishman S.
      • Kornblith A.B.
      • Kosty M.
      • Green M.R.
      • Holland J.
      Is quality of life predictive of the survival of patients with advanced nonsmall cell lung carcinoma?.
      using a similar design failed to show such a link. A study including 122 patients with Stage III and IV lung cancer patients also did not identify relationships between psychosocial factors and survival.
      • Akechi T.
      • Okamura H.
      • Okuyama T.
      • et al.
      Psychosocial factors and survival after diagnosis of inoperable non-small cell lung cancer.
      Faller et al.
      • Faller H.
      • Schmidt M.
      Prognostic value of depressive coping and depression in survival of lung cancer patients.
      could not confirm a relationship between depression and mortality but demonstrated the relevance of depressive coping for survival. In contrast, an epidemiologic study reported that psychiatric comorbidities decreased survival.
      • Dalton S.O.
      • Schüz J.
      • Engholm G.
      • et al.
      Social inequality in incidence of and survival from cancer in a population-based study in Denmark, 1994–2003: summary of findings.
      Yet given the epidemiologic nature of the study biomedical and treatment variables had not been controlled for. To date, no study controlling for biologic risk factors examined associations between anxiety and mortality in patients with non–small cell lung cancer, although anxiety is a highly prevalent symptom.
      Although anxiety and depression are correlated, they represent distinct symptom patterns, which should be studied separately. For example, anxiety can be a more volatile syndrome in cancer patients compared with depression,
      • Linden W.
      • MacKenzie R.
      • Rnic K.
      • Marshall C.
      • Vodermaier A.
      Emotional adjustment over 1 year post-diagnosis in patients with cancer: understanding and predicting adjustment trajectories.
      but this effect is not consistent.
      • Hernández Blázquez M.
      • Cruzado J.A.
      A longitudinal study on anxiety, depressive and adjustment disorder, suicide ideation and symptoms of emotional distress in patients with cancer undergoing radiotherapy.

      The Present Study

      Given the scarcity of studies on lung cancer and anxiety and inconsistency of findings with regard to depression, the present study tested associations between anxiety and depression on lung-cancer–specific and all-cause mortality in a defined sample of patients with Stage III non–small cell lung cancer awaiting upfront radiotherapy. We hypothesized that both anxiety and depression were associated with shorter length of survival, possibly because distressed lung cancer patients also tend to continue smoking despite their cancer diagnosis.
      • Cooley M.E.
      • Wang Q.
      • Johnson B.E.
      • et al.
      Factors associated with smoking abstinence among smokers and recent-quitters with lung and head and neck cancer.
      • Simmons V.N.
      • Litvin E.B.
      • Jacobsen P.B.
      • et al.
      Predictors of smoking relapse in patients with thoracic cancer or head and neck cancer.
      • Berg C.J.
      • Thomas A.N.
      • Mertens A.C.
      • et al.
      Correlates of continued smoking versus cessation among survivors of smoking-related cancers.

      Methods

      In 2004, a routine screening program for emotional distress was started, where every newly diagnosed cancer patient attending two major cancer centers in Metro Vancouver (BC Cancer Agency Vancouver and BC Cancer Agency Surrey) was asked to complete a psychological screening questionnaire, the Psychosocial Screen for Cancer (PSSCAN
      • Linden W.
      • Yi D.
      • Barroetavena M.C.
      • MacKenzie R.
      • Doll R.
      Development and validation of a psychosocial screening instrument for cancer.
      • Linden W.
      • Vodermaier A.
      • McKenzie R.
      • et al.
      The Psychosocial Screen for Cancer (PSSCAN): further validation and normative data.
      ), with resulting findings being added to the patient's chart. The study used a retrospective cohort design and collected additional information from the provincial Cancer Agency Information System (CAIS). Approval was obtained from the University of British Columbia–BC Cancer Agency (BCCA) Research Ethics Board.

      Participant Selection

      To maximize homogeneity and representativeness of the sample, Stage III non–small cell lung cancer (NSCLC) patients were selected as upfront radiotherapy (RT) is the predominant treatment modality in these patients and RT is only delivered at the BCCA within British Columbia. Thus, a high proportion of Stage III patients are referred to BCCA. In British Columbia, more than 80% of Stage III lung cancer patients are referred to BCCA for consideration of radiotherapy.
      In contrast, only a fraction of patients with Stage I and Stage II lung cancer are referred to BCCA because these rarely require RT, whereas a large proportion of these patients may have received adjuvant chemotherapy through community hospitals instead. Thus, inclusion of patients with early-stage lung cancer would have led to a significant selection bias. Likewise, only a fraction of Stage IV lung cancer patients are referred to the BCCA, and inclusion of this stage would have also increased the selection bias. All patients with Stage III NSCLC referred to two Greater Vancouver cancer centers between 2002 and 2010 were identified.
      Patients were excluded if PSSCAN scores in the chart were missing or incomplete. PSSCAN scores and other characteristics of interest were manually extracted from the patients' paper charts. These psychosocial variables were merged with electronically archived demographic and biomedical data from the CAIS system. The resulting database was then merged with censored data regarding the patients' survival status. Survival data were retrieved from BC Vital Statistics, the Provincial Vital Statistics Agency, who provides complete death lists for all deceased British Columbians on a monthly basis. Survival data had been extracted in May 2012.

      Nonpsychological Variables Collected

      Age, sex, marital status, ethnicity, employment status, performance status (measured by the Eastern Cooperative Oncology Group performance status), stage, histology, and the use of definitive surgery, radiation, and/or initial chemotherapy were abstracted from CAIS. The Eastern Cooperative Oncology Group performance status
      • Oken M.M.
      • Creech R.H.
      • Tormey D.C.
      • et al.
      Toxicity and response criteria of the Eastern Cooperative Oncology group.
      is a physician rating of patients' performance status. Its scores run from 0 to 5, with “0” denoting perfect health and “5” death.

      Distress Scores

      The 21-item PSSCAN
      • Linden W.
      • Yi D.
      • Barroetavena M.C.
      • MacKenzie R.
      • Doll R.
      Development and validation of a psychosocial screening instrument for cancer.
      • Linden W.
      • Vodermaier A.
      • McKenzie R.
      • et al.
      The Psychosocial Screen for Cancer (PSSCAN): further validation and normative data.
      assesses anxiety and depressive symptoms, perceived social support, desired social support, and quality of life via self-report. Only the anxiety and depression subscales were analyzed in this study. These consist of five questions each asking for the patients' level of anxiety and depression, respectively. The answer format is a five-point rating scale, with scores ranging from “Not at All” (Score 1) to “Very Much So” (Score 5). The sum of the five items of each subscale represents the subscale score.
      The PSSCAN was specifically developed for use with cancer patients, meets multiple criteria for reliability in cancer patients and healthy individuals, and also possesses content, concurrent, and construct validity.
      • Linden W.
      • Yi D.
      • Barroetavena M.C.
      • MacKenzie R.
      • Doll R.
      Development and validation of a psychosocial screening instrument for cancer.
      • Linden W.
      • Vodermaier A.
      • McKenzie R.
      • et al.
      The Psychosocial Screen for Cancer (PSSCAN): further validation and normative data.
      Internal consistency was α = 0.83 on average, and test-retest reliability over 2 months was r = 0.64.
      • Linden W.
      • Vodermaier A.
      • McKenzie R.
      • et al.
      The Psychosocial Screen for Cancer (PSSCAN): further validation and normative data.
      The validation process also included a control group of healthy individuals. Furthermore, comparisons of the PSSCAN anxiety and depression subscales against the extensively validated HADS scale
      • Vodermaier A.
      • Millman R.D.
      Accuracy of the Hospital Anxiety and Depression Scale as a screening tool in cancer patients: a systematic review and meta-analysis.
      demonstrated very good sensitivity (92% and 100%) and specificity (98% and 86%). Cutoffs for each subscale are 8 for subclinical symptoms and 11 for clinical symptoms.

      Statistical Methods

      To test the effect of anxiety and depression on mortality, Cox proportional hazards regression analysis was used for survival analysis adjusting for relevant demographic, biomedical, and treatment variables.
      The censor date, which is the most recent date for which death information was available, contains complete information on all patients in the registry, and all cases are censored at that same date. Censored cases were defined by nonreceipt of information about a patient's death and death date at the time of retrieval of survival data. Thus, patients for whom data on survival status were missing were categorized as censored. In this process, the censoring mechanism is not related to the outcome and the assumption of noninformative and right censoring is satisfied.
      Survival time was calculated by subtracting the date of death or censoring from the date of diagnosis. Analyses were conducted for death resulting from any cause (i.e., all-cause mortality) and death resulting from lung cancer (i.e., lung cancer–specific mortality).
      Cox proportional hazards regression analysis uses the likelihood ratio statistic and calculates the overall chi-square. Parameter estimates for each variable in the model represent hazard ratios. Analyses were controlling for the effects of age, sex, marital status, ethnicity, employment status, performance status, stage, histology, and treatment variables (i.e., surgery, radiation, chemotherapy).
      In a first step, proportionality of hazards was examined by testing the time dependency of each control or predictor variable separately. The respective variable was entered together with its interaction with time in a Cox regression model. If the variable × time interaction turned out to be significant, this respective variable demonstrates time-dependent changes in hazards, and a variable × time interaction, therefore, needs to be included in the multivariate model.
      • Bewick V.
      • Cheek L.
      • Ball J.
      Statistics review 12: survival analysis.
      Next, a forced entry hierarchical model was chosen where demographic variables were entered in a first step, biomedical (i.e., performance status, stage, histology) entered second, treatment variables (surgery, radiation, chemotherapy) entered third, and either anxiety or depression entered fourth.
      Age was entered as continuous, mean-centered variable into the Cox regression equation model. All other control variables were entered as categorical variables. Predictor variables were tested separately in its effects on survival and were entered as continuous variables because analyzing variables as continuous variables in contrast with analyzes of categorized variables is not associated with a loss of information.
      Significant effects were visualized by plotting Kaplan-Meier curves with the subclinical threshold used as the cutoff.
      All statistical tests were two-tailed and a P-value of less than 0.05 was considered statistically significant. Confidence intervals are also reported as an alternative indicator of significance, indicating that the population mean will fall within the respective range in 95% of samples. Statistical analyses were conducted with IBM SPSS Statistics, version 23.

      Results

      One thousand eighty patients were identified with Stage III NSCLC and 756 provided complete PSSCAN data. Missing biomedical and treatment data further reduced the sample number to 684 (To retain as many patients as possible for analyzes, categorical control variables included an “unknown” category. This strategy, however, cannot be applied to continuous variables.). Characteristics of the study cohort are presented in Table 1. Patients were 68 years on average and equally distributed by sex; 65.2% of the sample were married or lived in a common law relationship and 78.9% were white. Only 12.0% of patients were still in the work force. About two-third of patients had Stage IIIA disease. Among histologic subtypes, non–small cell carcinoma not otherwise specified (NOS) was with 43.7% the most prevalent, followed by squamous cell carcinoma (24.4%) and adenocarcinoma (23.4%). Only 15.8% of patients had undergone surgery, whereas 83.2% received upfront radiation treatment; 36.5% underwent upfront chemotherapy. Mean anxiety and depression scores reached the subclinical threshold. At the time of analysis, 553 of 684 subjects had died, of which 489 had died of lung cancer. Median overall survival was 12.5 months.
      Table 1Patient Characteristics
      VariablesTotal (n = 684)
      Age at diagnosis, mean (SD)68.22 (10.41)
      Sex
       Male341 (49.9%)
       Female343 (50.1%)
      Marital status
       Married/common law446 (65.2%)
       Single47 (6.9%)
       Divorced/separated81 (11.8%)
       Widowed109 (15.9%)
       Unknown11 (0.1%)
      Ethnicity
       Caucasian540 (78.9%)
       Asian108 (15.8%)
       Other36 (5.3%)
      Employment status
       Working82 (12.0%)
       Not working602 (88.0%)
      ECOG performance status
       065 (9.5%)
       1250 (36.5%)
       2128 (18.7%)
       3128 (18.7%)
       49 (1.3%)
       Unknown104 (15.2%)
      Stage
       IIIA446 (65.2%)
       IIIB238 (34.8%)
      Histology
       Non–small cell not otherwise specified299 (43.7%)
       Adenocarcinoma160 (23.4%)
       Squamous cell167 (24.4%)
       Large cell16 (2.3%)
       Bronchoalveolar carcinoma2 (.3%)
       Other40 (5.8%)
      Surgery
       Yes108 (15.8%)
       No572 (83.6%)
       Unknown4 (.6%)
      Radiation
       Yes569 (83.2%)
       No115 (16.8%)
      Chemotherapy
       Yes250 (36.5%)
       No432 (63.2%)
       Unknown2 (.3%)
      Anxiety, mean (SD)8.87 (4.15)
      Anxiety (subclinical threshold)
       Anxious198 (26.2%)
       Not anxious360 (47.6%)
      Depression, mean (SD)7.96 (3.62)
      Depression (subclinical threshold)
       Depressed148 (39.6%)
       Not depressed457 (60.4%)
      All-cause mortality
       All553 (80.8%)
       Depressed234 (85.7%)
       Anxious295 (82.4%)
      Lung cancer mortality
       All489 (71.5%)
       Depressed210 (76.9%)
       Anxious256 (71.5%)
      Median length of survival (months)
       All12.47
       Depressed10.78
       Anxious11.32
      Median length of follow-up (months)64.50
      ECOG = Eastern Cooperative Oncology Group.
      Anxiety and depression were moderately to strongly positively correlated at ρ = 0.64. Anxiety and depression were also modestly positively correlated with performance status (ρ = 0.11 for anxiety; ρ = 0.20, for depression). Age at diagnosis exhibited small negative correlations with anxiety (ρ = −0.14) and depression (ρ = −0.10) and was positively associated with performance status (ρ = 0.18).
      Unadjusted analyses showed both significant associations between anxiety and all-cause mortality (χ2(1) = 4.31; P = 0.038) and between depression and lung cancer–specific (χ2(1) = 8.79; P = 0.003) and all-cause mortality (χ2(1) = 7.38; P = 0.007). Kaplan-Meier curves of these unadjusted effects are displayed in Fig. 1a and 1b for anxiety and in Fig. 2a and 2b for depression, respectively.
      Figure thumbnail gr1
      Fig. 1a) Anxiety and lung cancer–specific survival. b) Anxiety and overall survival.
      Figure thumbnail gr2
      Fig. 2a) Depression and lung cancer–specific survival. b) Depression and overall survival.
      Adjusted analyses demonstrated significant associations between anxiety on lung cancer–specific (hazard ratio [HR] 1.04; 95% confidence interval [CI] 1.01–1.07; P = 0.035) and all-cause mortality (HR 1.04; 95% CI 1.01–1.07; P = 0.005) (Table 2), whereas no significant effect of depression on mortality emerged (Table 3). Accounting for the significant intercorrelation of depression with performance status and a potential confounder effect of performance status on the outcome, analyzes were rerun while omitting performance status as a predictor. As a result, depression was significantly associated with lung cancer–specific (HR 1.03; 95% CI 1.01–1.06; P = 0.008) and all-cause mortality (HR 1.04; 95% CI 1.01–1.06; P = 0.003).
      Table 2Results of Cox Proportional-Hazards Regression Analyses: Anxiety
      VariablesLung Cancer MortalityAll-Cause Mortality
      HR (95% CI)P-valueHR (95% CI)P-value
      Age1.00 (0.99–1.01)0.9411.00 (0.99–1.01)0.541
      Sex1.28 (1.05–1.55)0.0151.35 (1.12–1.62)0.001
      Marital status
       Married/common law10.29410.305
       Single/divorced/widowed0.86 (0.70–1.05)0.1450.86 (0.71–1.05)0.139
       Unknown1.54 (0.20–11.71)0.6751.35 (0.18–10.19)0.770
      Employment status0.98 (0.71–1.33)0.8770.95 (0.71–1.28)0.752
      Ethnicity
       Caucasian10.51110.534
       Asian0.87 (0.67–1.14)0.3080.87 (0.68–1.12)0.278
       Other0.87 (0.56–1.34)0.5180.93 (0.62–1.38)0.702
      ECOG performance status
       01<0.0011<0.001
       10.59 (0.38–0.90)0.0150.62 (0.41–0.92)0.019
       20.79 (0.59–1.04)0.0950.80 (0.61–1.05)0.105
       30.99 (0.72–1.35)0.9411.02 (0.76–1.37)0.901
       41.51 (1.11–2.06)0.0091.56 (1.16–2.09)0.003
       Unknown2.35 (1.14–4.85)0.0212.18 (1.06–4.47)0.033
      Stage1.28 (1.05–1.55)0.0131.23 (1.02–1.47)0.027
      Histology
       Non-small cell lung cancer NOS10.30710.224
       Adenocarcinoma1.11 (0.88–1.42)0.3781.13 (0.90–1.42)0.288
       Squamous cell1.20 (0.95–1.51)0.1211.21 (0.98–1.50)0.083
       Large cell1.09 (0.60–1.20)0.7740.96 (0.53–1.74)0.903
       BAC3.08 (0.73–13.05)0.1262.82 (0.67–11.84)0.157
       Other0.86 (0.57–1.29)0.4570.84 (0.57–1.24)0.376
      Surgery
       Yes1<0.0011<0.001
       No0.41 (0.29–0.59)<0.0010.42 (0.30–0.58)<0.001
       Unknown1.39 (0.33–5.85)0.6561.17 (0.28–4.92)0.827
      Radiology0.66 (0.43–1.02).0620.68 (0.46–1.02)0.062
      Chemotherapy
       Yes1<0.00110.001
       No0.48 (0.33–0.67)<0.0010.54 (0.39–0.75)<0.001
       Unknown0.16 (0.00–6.49)0.3310.18 (0.01–7.02)0.357
      Anxiety1.04 (1.01–1.07)0.0351.04 (1.01–1.07)0.005
      Radiology × time1.04 (1.01–1.07)0.0081.04 (1.01–1.07)0.005
      Chemotherapy × time10.04810.134
       11.02 (1.01–1.04)0.0141.02 (1.00–1.03)0.047
       21.04 (0.86–1.25)0.6861.03 (0.86–1.24)0.726
      Anxiety × time1.00 (1.00–1.00)0.0331.00 (1.00–1.00)0.039
      CI = confidence interval; HR = hazard ratio; ECOG = Eastern Cooperative Oncology Group.
      Test statistics represent hazard ratios with 95% confidence interval (HR [95% CI]).
      Table 3Results of Cox Proportional-Hazards Regression Analyses: Depression
      VariablesLung Cancer MortalityAll-Cause Mortality
      HR (95% CI)P-valueHR (95% CI)P-value
      Age1.00 (0.99–1.01)0.8821.00 (0.99–1.01)0.526
      Sex1.28 (1.05–1.56)0.0131.34 (1.12–1.62)0.002
      Marital status
       Married/common law10.36710.357
       Single/divorced/widowed0.87 (0.71–1.07)0.1800.87 (0.72–1.06)0.162
       Unknown1.42 (0.19–10.80)0.7361.24 (0.16–9.38)0.834
      Employment status0.97 (0.71–1.33)0.8540.95 (0.71–1.28)0.757
      Ethnicity
       Caucasian10.52110.519
       Asian0.87 (0.67–1.14)0.3010.87 (0.67–1.12)0.267
       Other0.88 (0.57–1.36)0.5650.93 (0.62–1.38)0.706
      ECOG performance status
       01<0.0011<0.001
       10.59 (0.39–0.91)0.0170.63 (0.42–0.94)0.024
       20.78 (0.59–1.04)0.0890.80 (0.61–1.04)0.099
       30.96 (0.70–1.32)0.8211.00 (0.75–1.35)0.981
       41.53 (1.12–2.08)0.0071.59 (1.19–2.14)0.002
       Unknown2.50 (1.22–5.13)0.0132.36 (1.16–4.82)0.018
      Stage1.28 (1.05–1.55)0.0131.23 (1.02–1.47)0.029
      Histology
       Non-small cell lung cancer NOS10.34810.271
       Adenocarcinoma1.11 (0.87–1.41)0.4061.12 (0.90–1.41)0.310
       Squamous cell1.20 (0.95–1.50)0.1291.20 (0.97–1.49)0.091
       Large cell1.11 (0.61–2.02)0.7230.99 (0.55–1.78)0.961
       BAC2.88 (0.68–12.19)0.1502.64 (0.63–11.08)0.185
       Other0.86 (0.57–1.30)0.4640.84 (0.57–1.25)0.395
      Surgery
       Yes1<0.0011<0.001
       No0.41 (0.29–0.59)<0.0010.43 (0.31–0.59)<0.001
       Unknown1.43 (0.34–6.04)0.6241.18 (0.28–4.92)0.825
      Radiology0.68 (0.44–1.05).0840.70 (0.47–1.06)0.090
      Chemotherapy
       Yes1<0.00110.001
       No0.48 (0.34–0.68)<0.0010.55 (0.40–0.75)<0.001
       Unknown0.14 (0.00–5.77)0.3020.17 (0.00–6.45)0.336
      Depression1.02 (0.99–1.05)0.1571.02 (0.99–1.05)0.133
      Radiology × time1.04 (1.01–1.07)0.0111.04 (1.01–1.07)0.007
      Chemotherapy × time10.05310.140
       11.02 (1.00–1.04)0.0171.02 (1.00–1.03)0.050
       21.04 (0.87–1.25)0.6501.04 (0.87–1.24)0.696
      CI = confidence interval; HR = hazard ratio; ECOG = Eastern Cooperative Oncology Group.
      Test statistics represent hazard ratios with 95% confidence interval (HR [95% CI]).

      Discussion

      Patients with Stage III NSCLC who reported greater anxiety after diagnosis had a greater hazard of lung cancer–specific and all-cause mortality at follow-up. In unadjusted analyses, both anxiety and depression predicted shorter length of survival. However, associations between depression after diagnosis and mortality did not reach significance in multivariate analyses, whereas the effect of anxiety on mortality remained. Given that depression was more strongly correlated with performance status than anxiety, this evident discrepancy raises the question of overlapping variances. The fact that a significant effect emerged again after exclusion of performance status from analyses indicates a potential confounder effect of performance status. Because relations between depression and performance status are bidirectional, a mediational pathway cannot be presumed: patients with poor performance status often report functional limitations and pain and, thus, are more likely to feel depressed. Conversely, depressed patients likewise report functional impairment, which can result as a consequence of reduced energy and of cognitive distortions, both core symptoms of depression.
      Although widely examined in other cancer types, studies that tested associations between emotional distress and mortality in patients with lung cancer are scarce to date, and many of these are based on nonrepresentative small sample sizes and lack biomedical and treatment prognosticators. The present study, therefore, provides evidence that anxiety and eventually depression influence length of survival in advanced lung cancer. Nevertheless, the effects observed are small. It has been hypothesized previously that psychosocial factors may not play a role in more aggressive types of cancer,
      • Levy S.M.
      • Roberts D.C.
      Psychoneuroimmunology: prediction of cancer outcomes.
      but this observation may have arisen from studies with smaller samples and less statistical power.
      Limitations of the study include the lack of information about smoking habits after lung cancer, a well-known moderator of mortality in this patient group and on biochemical measures (albumin, lactate dehydrogenase), weight loss, and mediastinal node (N2) status,
      • Espinosa E.
      • Feliu J.
      • Zamora P.
      • et al.
      Serum albumin and other prognostic factors related to response and survival in patients with advanced non-small cell lung cancer.
      • Jafri S.H.
      • Shi R.
      • Mills G.
      Advance lung cancer inflammation index (ALI) at diagnosis is a prognostic marker in patients with metastatic non-small cell lung cancer (NSCLC): a retrospective review.
      yet the respective relevance of some prognostic factors remains unclear for patients with Stage III NSCLC.
      • Berghmans T.
      • Paesmans M.
      • Sculier J.P.
      Prognostic factors in stage III non-small cell lung cancer: a review of conventional, metabolic and new biological variables.
      Ideally, anxiety and depression could have been assessed via use of Structured Clinical Interviews, which compared with questionnaire assessments reflect a superior method. A disadvantage of interview assessments are a) its time intensity, which is difficult to establish in routine medical care, where the data of this study had been collected, and b) it would have omitted the consideration of subclinical symptom levels, which can involve dysfunctional health behavior including continued smoking, potentially influencing the course of the disease. Furthermore, anxiety and depression scores were only recorded once, and the potential impact of varying trajectories of distress and its effects on survival, therefore, could not be investigated. Similarly, physician-rated performance status was only assessed once and is likewise a construct whose parameter values can change across time.
      Despite these limitations, this study's strengths include a large, representative patient sample within a well-defined tumor type and stage, allowing for accurate control of many relevant biomedical prognostic factors.
      No information was available on whether lung cancer patients continued smoking after diagnosis. It is well known that a significant proportion of lung cancer patients continue smoking or fail smoking cessation attempts.
      • Sanderson Cox L.
      • Sloan J.A.
      • Patten C.A.
      • et al.
      Smoking behavior of 226 patients with diagnosis of stage IIIA/IIIB non-small cell lung cancer.
      • Tsao A.S.
      • Liu D.
      • Lee J.J.
      • Spitz M.
      • Hong W.K.
      Smoking affects treatment outcome in patients with advanced nonsmall cell lung cancer.
      Emotional distress and low self-efficacy are interconnected therewith.
      • Schnoll R.A.
      • Malstrom M.
      • James C.
      • et al.
      Correlates of tobacco use among smokers and recent quitters diagnosed with cancer.
      • Schnoll R.A.
      • James C.
      • Malstrom M.
      • et al.
      Longitudinal predictors of continued tobacco use among patients diagnosed with cancer.
      Consequently, psychological interventions should target these correlates of continued smoking to ultimately improve patients' quality of life and eventual length of survival.

      Disclosures and Acknowledgments

      This research received no specific funding/grant from any funding agency in the public, commercial, or not-for-profit sectors. The authors declare no conflicts of interest.

      References

        • Govindan R.
        • Page N.
        • Morgensztern D.
        • et al.
        Changing epidemiology of small-cell lung cancer in the United States over the last 30 years: analysis of the surveillance, epidemiologic, and end results database.
        J Clin Oncol. 2006; 24: 4539-4544
        • Siegel R.L.
        • Miller K.D.
        • Jemal A.
        Cancer statistics, 2015.
        CA Cancer J Clin. 2015; 65: 5-29
        • Alberg A.J.
        • Brock M.V.
        • Ford J.G.
        • Samet J.M.
        • Spivack S.D.
        Epidemiology of lung cancer: diagnosis and management of lung cancer, 3rd ed.: American College of Chest Physicians evidence-based clinical practice guidelines.
        Chest. 2013; 143: 1S-29S
        • de Groot P.
        • Munden R.F.
        Lung cancer epidemiology, risk factors, and prevention.
        Radiol Clin North Am. 2012; 50: 863-876
        • Walters S.
        • Maringe C.
        • Coleman M.P.
        • et al.
        Lung cancer survival and stage at diagnosis in Australia, Canada, Denmark, Norway, Sweden and the UK: a population-based study, 2004-2007.
        Thorax. 2013; 68: 551-564
        • Linden W.
        • Vodermaier A.
        • Mackenzie R.
        • Greig D.
        Anxiety and depression after cancer diagnosis: prevalence rates by cancer type, gender, and age.
        J Affect Disord. 2012; 141: 343-351
        • Vodermaier A.
        • Linden W.
        • MacKenzie R.
        • Greig D.
        • Marshall C.
        Disease stage predicts post-diagnosis anxiety and depression only in some types of cancer.
        Br J Cancer. 2011; 105: 1814-1817
        • Chambers S.K.
        • Dunn J.
        • Occhipinti S.
        • et al.
        A systematic review of the impact of stigma and nihilism on lung cancer outcomes.
        BMC Cancer. 2012; 12: 184
        • Ader R.
        • Cohen N.
        • Felten D.
        Psychoneuroimmunology: interactions between the nervous system and the immune system.
        Lancet. 1995; 345: 99-103
        • Reiche E.M.
        • Nunes S.O.
        • Morimoto H.K.
        Stress, depression, the immune system, and cancer.
        Lancet Oncol. 2004; 5: 617-625
        • Ganz P.A.
        • Lee J.J.
        • Siau J.
        Quality of life assessment. An independent prognostic variable for survival in lung cancer.
        Cancer. 1991; 67: 3131-3135
        • Herndon J.E.
        • Fleishman S.
        • Kornblith A.B.
        • Kosty M.
        • Green M.R.
        • Holland J.
        Is quality of life predictive of the survival of patients with advanced nonsmall cell lung carcinoma?.
        Cancer. 1999; 85: 333-340
        • Akechi T.
        • Okamura H.
        • Okuyama T.
        • et al.
        Psychosocial factors and survival after diagnosis of inoperable non-small cell lung cancer.
        Psychooncology. 2009; 18: 23-29
        • Faller H.
        • Schmidt M.
        Prognostic value of depressive coping and depression in survival of lung cancer patients.
        Psychooncology. 2004; 13: 359-363
        • Dalton S.O.
        • Schüz J.
        • Engholm G.
        • et al.
        Social inequality in incidence of and survival from cancer in a population-based study in Denmark, 1994–2003: summary of findings.
        Eur J Cancer. 2008; 44: 2074-2085
        • Linden W.
        • MacKenzie R.
        • Rnic K.
        • Marshall C.
        • Vodermaier A.
        Emotional adjustment over 1 year post-diagnosis in patients with cancer: understanding and predicting adjustment trajectories.
        Support Care Cancer. 2015; 23: 1391-1399
        • Hernández Blázquez M.
        • Cruzado J.A.
        A longitudinal study on anxiety, depressive and adjustment disorder, suicide ideation and symptoms of emotional distress in patients with cancer undergoing radiotherapy.
        J Psychosom Res. 2016; 87: 14-21
        • Cooley M.E.
        • Wang Q.
        • Johnson B.E.
        • et al.
        Factors associated with smoking abstinence among smokers and recent-quitters with lung and head and neck cancer.
        Lung Cancer. 2012; 76: 144-149
        • Simmons V.N.
        • Litvin E.B.
        • Jacobsen P.B.
        • et al.
        Predictors of smoking relapse in patients with thoracic cancer or head and neck cancer.
        Cancer. 2013; 119: 1420-1427
        • Berg C.J.
        • Thomas A.N.
        • Mertens A.C.
        • et al.
        Correlates of continued smoking versus cessation among survivors of smoking-related cancers.
        Psychooncology. 2013; 22: 799-806
        • Linden W.
        • Yi D.
        • Barroetavena M.C.
        • MacKenzie R.
        • Doll R.
        Development and validation of a psychosocial screening instrument for cancer.
        Health Qual Life Outcomes. 2005; 3: 54
        • Linden W.
        • Vodermaier A.
        • McKenzie R.
        • et al.
        The Psychosocial Screen for Cancer (PSSCAN): further validation and normative data.
        Health Qual Life Outcomes. 2009; 7: 16
      1. BC Cancer Registry Annual Report. BC Cancer Agency, Vancouver, Canada2013
        • Oken M.M.
        • Creech R.H.
        • Tormey D.C.
        • et al.
        Toxicity and response criteria of the Eastern Cooperative Oncology group.
        Am J Clin Oncol. 1982; 5: 649-655
        • Vodermaier A.
        • Millman R.D.
        Accuracy of the Hospital Anxiety and Depression Scale as a screening tool in cancer patients: a systematic review and meta-analysis.
        Support Care Cancer. 2011; 19: 1899-1908
        • Bewick V.
        • Cheek L.
        • Ball J.
        Statistics review 12: survival analysis.
        Crit Care. 2004; 8: 389-394
        • Levy S.M.
        • Roberts D.C.
        Psychoneuroimmunology: prediction of cancer outcomes.
        in: Schneider N. McCabe P. Baum A. Behavioral medicine: Stress and disease processes. Erlbaum, Hillsdale1992: 165-174
        • Espinosa E.
        • Feliu J.
        • Zamora P.
        • et al.
        Serum albumin and other prognostic factors related to response and survival in patients with advanced non-small cell lung cancer.
        Lung Cancer. 1995; 12: 67-76
        • Jafri S.H.
        • Shi R.
        • Mills G.
        Advance lung cancer inflammation index (ALI) at diagnosis is a prognostic marker in patients with metastatic non-small cell lung cancer (NSCLC): a retrospective review.
        BMC Cancer. 2013; 13: 158
        • Berghmans T.
        • Paesmans M.
        • Sculier J.P.
        Prognostic factors in stage III non-small cell lung cancer: a review of conventional, metabolic and new biological variables.
        Ther Adv Med Oncol. 2011; 3: 127-138
        • Sanderson Cox L.
        • Sloan J.A.
        • Patten C.A.
        • et al.
        Smoking behavior of 226 patients with diagnosis of stage IIIA/IIIB non-small cell lung cancer.
        Psychooncology. 2002; 11: 472-478
        • Tsao A.S.
        • Liu D.
        • Lee J.J.
        • Spitz M.
        • Hong W.K.
        Smoking affects treatment outcome in patients with advanced nonsmall cell lung cancer.
        Cancer. 2006; 106: 2428-2436
        • Schnoll R.A.
        • Malstrom M.
        • James C.
        • et al.
        Correlates of tobacco use among smokers and recent quitters diagnosed with cancer.
        Patient Educ Couns. 2002; 46: 137-145
        • Schnoll R.A.
        • James C.
        • Malstrom M.
        • et al.
        Longitudinal predictors of continued tobacco use among patients diagnosed with cancer.
        Ann Behav Med. 2003; 25: 214-222