You are here
The risk of being depressed is significantly higher in cancer patients than in the general population: Prevalence and severity of depressive symptoms across major cancer types
European Journal of Cancer, Volume 72, February 2017, Pages 46-53
Depression is a common co-morbidity of cancer that has a detrimental effect on quality of life, treatment adherence and potentially survival. We conducted an epidemiological multi-center study including a population-based random comparison sample and estimated the prevalence of depressive symptoms by cancer site, thereby identifying cancer patients with the highest prevalence of depression.
Patients and methods
We included 4020 adult cancer inpatients and outpatients from five distinct regions across Germany in a proportional stratified random sample based on the nationwide cancer incidence and a comparison group consisting of 5018 participants. Both groups reported depressive symptoms by filling in the Patient Health Questionnaire (PHQ-9). In multivariate analyses adjusted for age and sex, we calculated the odds of being depressed.
Out of 5818 eligible patients, 69% participated (51% women, mean age = 58 years). We estimated that one in four cancer patients (24%) is depressed (PHQ-9 ≥ 10). The odds of being depressed among cancer patients were more than five times higher than in the general population (OR, 5.4; 95% CI, 4.6–6.2). Patients with pancreatic (M = 8.0, SD = 5.0), thyroid (M = 7.8, SD = 6.3) and brain tumours (M = 7.6, SD = 4.9) showed the highest prevalence, whereas patients with prostate cancer (M = 4.3, SD = 3.8) and malignant melanoma (M = 5.3, SD = 4.3) had the lowest levels of depressive symptoms.
Our results help clinicians identify cancer patients in need of psychosocial support when navigating in the growing survivor population.
- One in four cancer patients is depressed.
- The odds of being depressed are five times higher in cancer patients.
- People with brain tumours and thyroid cancer were the most depressed.
Keywords: Cancer, Depression, Psychosocial oncology, Epidemiology, PHQ-9.
Clinical depression is a common immediate and late co-morbidity of cancer with prevalence estimates ranging from 4% to 16% over the first 5 years following diagnosis  and . Beyond suffering from the symptoms of depression itself, depressed cancer patients show lower treatment adherence and even worse survival compared with non-depressed cancer patients . Oncology clinicians as well as GPs could improve both quality of life and survival in cancer patients by adequately diagnosing and treating depression.
Patients with different types of cancer have been known to vary largely in their risk of suffering from depression. Yet there have been few, if any, comprehensive epidemiological studies comparing depressive symptoms across all major cancer types and treatment settings. Clinical depression can be assessed by diagnostic interviews  or inferred from register-based data such as psychiatric hospitalisations  or prescriptions of anti-depressants . We chose to use self-reported depressive symptoms as these also capture sub-threshold symptom burden, which may be due to paraneoplastic syndromes or cancer treatment but may still need to be treated without a clinical diagnosis of major depression.
One of the largest self-report studies published so far (N = 1, 385), which used the PHQ-4, a version of the Patient Health Questionnaire (PHQ), observed the highest depressive symptom burden in patients with cancers of the female genital organs (30% depressed), followed by lung (22%) and breast cancer (21%); patients with malignant melanoma reported the lowest rates of depressive symptoms in their study (8%) . However, this article neither reported the cut-off value that was used nor data on cancers of the urinary tract, haematological malignancies and thyroid cancer.
Meta-analyses have managed to cover a broader range of cancer types and settings  and . Across these studies, patients with cancers of the female genital organs, lung and pancreatic (digestive tract) cancers showed high rates of depressive symptoms, but overall results were heterogeneous.
However, these meta-analyses summarised patients in heterogeneous groups with largely varying levels of symptom burden, such as Krebber et al. who did not differentiate between different cancers of the digestive tract, e.g. pancreatic or colorectal cancer. In addition, these analyses have combined studies using either relatively liberal or more conservative instruments and as a result some groups may have been overestimated or underestimated in their prevalence of depression. Furthermore, data on rare cancers such as brain tumours or thyroid cancer were sparse or completely absent.
To identify those cancer patients who are most depressed, we conducted an epidemiological multi-center study, randomly sampling cancer patients to ensure a reliable comparison of depressive symptoms across patients diagnosed with and treated for all types of cancer, including rare cancer types. We compared depressive symptoms in cancer patients with a large random population-based sample, adjusting for age and sex.
2.1. Study protocol
The full study protocol to this study has been published elsewhere  but briefly followed these steps:
2.2. Patients and procedures
We used a proportional stratified random sample based on the nationwide incidence of all cancer diagnoses in Germany . Patients were consecutively recruited from a total of 84 inpatient oncology wards, outpatient oncology clinics and cancer rehabilitation centres in five distinct regions across Germany (Freiburg, Hamburg, Heidelberg, Leipzig and Würzburg).
Inclusion criteria were a confirmed diagnosis of a malignant tumour, age 18 through 75 and proficiency in German. Patients were excluded if they showed cognitive or verbal impairments that interfered with their ability to give informed consent for research. The study received research ethics committee approval by Medical Associations in all federal states involved (File reference numbers: Hamburg: 2768; Schleswig–Holstein: 61/09; Freiburg: 244/07; Heidelberg: S-228/2007-50155039; Würzburg: 107/07; and Leipzig: 200–2007). All participants provided written informed consent following information given both orally and in writing according to the Declaration of Helsinki. Data protection was secured in accordance with the German data protection laws (§§27-30a BDSG ).
2.3. General population data
To compare depressive symptoms between cancer patients and the general population, we used a large population-based sample (N = 5018). Two cohorts were randomly drawn from the adult German population living in private households in 2003 (N = 2500) and 2008 (N = 2518), using the random route technique based on the register of the Federal Elections: random selection of areas, random selection of households within these areas and random selection of individuals within those households. The total response rate was 63%. The sample was representative in terms of age (M = 49 years), sex (54% women) and education. Sample characteristics of this group were previously reported in more detail .
Demographic characteristics obtained were sex, age, partnership and years of education. Medical information was obtained from medical records and comprised cancer diagnosis (ICD-10 ), date of current cancer diagnosis, tumour stage (UICC TNM classification ), treatment intention (curative/palliative), completed treatments (surgery, radiation, chemotherapy, hormone treatment) and somatic co-morbidities.
Depressive symptoms were assessed with the German version of the PHQ-9, which is a validated, reliable and widely used self-report measure , , and . The PHQ-9 is a 9-item depression module from the full PHQ, scoring each of the 9 DSM-IV criteria of major depressive disorder as ‘0’ (not at all) to ‘3’ (nearly every day) over the previous two weeks: (1) anhedonia, (2) depressed mood, (3) trouble sleeping, (4) feeling tired, (5) change in appetite, (6) guilt or worthlessness, (7) trouble concentrating, (8) feeling slowed down or restless, (9) suicidal thoughts. The total PHQ-9 score can thus range from 0 to 27. The following cut-off scores have been recommended to determine levels of depression severity: 0–4 minimal, 5–9 mild, 10–14 moderate, 15–19 moderately severe, and 20–27 severe . To dichotomise scores into ‘depressed’ and ‘not depressed’, we used the most consistently recommended cut point of ≥10 .
The ECOG Scale was used to assess patients' performance status, i.e. daily living abilities . A person is evaluated on a score ranging from 0 (asymptomatic) to 4 (bedbound).
2.5. Statistical analysis
We performed descriptive, comparative and multivariate statistical analyses using the Statistical Package for the Social Sciences (SPSS; IBM, Armonk, NY), version 20.
Participants and non-responders were compared in multiple logistic regression models to calculate and test adjusted differences in age, sex, education, setting, study center or cancer type. We compared patient groups from different treatment settings in terms of age and sex in a one-way analysis of variance (ANOVA).
Sociodemographic and medical groups were compared in terms of depressive symptoms (PHQ-9 mean scores) with one-way ANOVA. The association between time since diagnosis as well as somatic co-morbidities and severity of depressive symptoms was tested with Pearson's r.
To compare patients with different cancer types and from different treatment settings to the general population, we performed multivariate analyses using the logistic regression feature in SPSS. We adjusted odds ratios for confounders by including age and sex in the regression model. To avoid variance attribution based on chance, age and sex were eliminated from the final regression model if these factors did not account for a statistically significant portion of the variance (P ≥ 0.05).
3.1. Non-responder analyses
Out of 5818 eligible patients, 4020 (69%) participated in the study. Most frequent reasons for non-participation were ‘too burdensome’ (N = 588; 33%) and organisational barriers (N = 111; 6%). As reported previously for this sample , non-responder analyses revealed that study participants were younger (P < 0.001), more educated (P < 0.001) and were more likely to be recruited from a cancer rehabilitation center (P < 0.001) than non-participants (Appendix Table A1). Participants and non-responders did not differ in terms of sex ratio (P = 0.096).
Sample characteristics are shown in Table 1. Patients' mean age was 58 years (SD = 11 years) and mean time since current diagnosis was 14 months (SD = 25). Patients whose tumour stage could not be determined were mostly patients with stage III or stage IV tumours and unclear metastatic status. Cancer types summarised as ‘other’ included mesothelioma (ICD-10: C45; N = 22, 14%), bone (C40–41; N = 17, 10%) and skin cancers other than malignant melanoma (C44; N = 13, 8%), thymus (C37, N = 8, 5%) and penile cancer (C60; N = 4, 3%). In terms of treatment setting, patients from inpatient oncology wards had the highest (56% men; M = 59 years) and patients from outpatient clinics had the lowest proportion of men (41% men; M = 59 years), whereas patients from cancer rehabilitation were younger (M = 57 years; 54% women; one-way ANOVA, P < 0.001).
Sociodemographic and medical sample characteristics and levels of depressive symptoms (PHQ-9) in cancer patients and the general population in Germany.
|Total sample||Depressed (PHQ-9 cut-off ≥10)||PHQ-9|
|Age category a||<0.001|
|In a relationship a||<0.001|
|Work situation a||<0.001|
|Cancer care setting||<0.001|
|Inpatient oncology wards||1735||43||410||24||6.4||4.8|
|Outpatient oncology clinics||1324||33||279||21||6.1||4.6|
|Cancer type (ICD-10)||<0.001|
|Female genital organs (C51–57)||317||8||92||29||7.5||4.8|
|Head and neck (C02–15, C32)||127||3||33||26||6.6||4.8|
|Soft tissue (C49)||39||1||10||26||7.4||5.0|
|Kidney/urinary tract (C64–66)||131||3||31||24||6.3||5.1|
|Malignant melanoma (C43)||67||2||11||16||5.3||4.3|
|Current disease condition a||<0.001|
|Not in remission||2365||61||600||25||6.7||4.8|
|Tumour stage (UICC TNM) a||<0.001|
|Treatment intention a||<0.001|
|Hormone therapy a||0.384|
|ECOG performance status a||<0.001|
a Reduced sample size due to missing data; P-values based on one-way ANOVA.
Abbreviation: PHQ-9, depression module of the Patient Health Questionnaire.
3.3. Depressive symptoms in cancer patients
As illustrated in Table 2, multivariate logistic regression analyses adjusted for age and sex revealed that cancer patients of all cancer types and treatment settings had significantly higher odds of being depressed than the general population. Patients with brain tumours and thyroid cancer were most likely to be depressed with their odds more than 9 times higher than those in the general population, while the lowest odds were found in patients with prostate cancer.
Odds ratios for risk of depression in cancer patients (N = 4020) compared to a general population sample (N = 5018) in Germany.
|Depressed (PHQ-9 ≥ 10)|
|All cancer patients||5.4b||4.6||6.2||<0.001|
|Inpatient oncology wards||5.4b||4.6||6.4||<0.001|
|Outpatient oncology clinics||4.6||3.8||5.4||<0.001|
|Female genital organs||7.2||5.4||9.7||<0.001|
|Head and neck||5.5a||3.6||8.3||<0.001|
Significant confounders adjusted for: aage, bsex.
Abbreviations: OR, odds ratio; PHQ-9, depression module of the Patient Health Questionnaire; 95% CI, 95% confidence interval; LB, lower boundary; UB, upper boundary.
On average, 24% of patients showed elevated depressive symptoms (PHQ-9 ≥ 10), whereas more severe symptoms (PHQ ≥ 15) were relatively rare across cancer types (Fig. 1).
Severity of depressive symptoms (PHQ-9 scores) by cancer type in Germany. PHQ-9, Patient Health Questionnaire.
The highest prevalence of depression was observed in middle aged, unemployed or single patients, patients in cancer rehabilitation, patients who had received chemotherapy and patients who had been diagnosed with metastatic and/or stage IV cancer (Table 1). Time since current cancer diagnosis showed a small but significant positive correlation with higher levels of depressive symptoms (r = 0.06, P < 0.001). The number of somatic co-morbidities was not significantly associated with depressive symptoms. We found small inverse associations between depressive symptoms and hypertension (r = −0.11, P < 0.001) and eye diseases including glaucoma (r = −0.03, P < 0.05). In addition, fatigue (r = 0.04, P < 0.05) and chronic inflammatory bowel diseases (r = 0.03, P < 0.05) were positively associated with depressive symptoms (Table A2 [available online]).
We found that cancer patients have five-fold increased odds of being depressed compared with the general population. Our findings arise from a large epidemiological multi-center study comparing depressive symptoms in cancer patients with a large population-based comparison group. This enabled us not only to detect the high depressive symptom burden in patients with thyroid, brain and pancreatic cancers but also to reliably compare them to other cancer patients as well as the general population.
4.1. Prevalence estimates compared with previous findings
Our prevalence estimates are comparable to those in a meta-analysis by Krebber et al. and cancer types ranked in approximately the same order, except for the three most severely depressed patient groups in our sample: thyroid, brain and pancreatic cancer, for whom data were absent from their study.
Interview-based studies have generally reported a lower prevalence of depression than studies using self-report instruments. This was also the case in a recent study with a subset of our sample using the CIDI-O interview (4-week prevalence of any mood disorder including major depressive and dysthymic disorder: 7%, N = 2141; ). Some of the patients with elevated, but sub-threshold depressive symptoms may have been diagnosed with adjustment disorders (4-week prevalence 11%; ); in others, elevated levels of depressive symptoms may be explained by the pathophysiological mechanisms outlined below.
4.2. Mechanisms underlying variations in depressive symptoms
Different mechanisms may account for the variations in depressive symptoms we observed. First, being diagnosed with and treated for a life-threatening disease may be seen as a severe life-event and depressive symptoms may be a reaction to this event. In addition, depressive symptoms may arise from cancer-related psychosocial factors, including existential distress or loss of meaning, changes in social role functioning or perceived body image, helplessness and hopelessness in the face of an uncontrollable situation , , and . According to this model, depressive symptoms would be expected to be associated with worse prognosis, more severe symptoms and treatment side-effects. Indeed, a diagnosis of a cancer with a particularly bad prognosis such as pancreatic cancer and brain tumours were strongly associated with elevated depressive symptoms in our sample. However, thyroid cancer patients were most depressed in our sample, although this cancer has a relatively good prognosis.
Second, cancer and its treatment may directly cause depressive symptoms through identifiable pathophysiological mechanisms. Thyroid cancer is typically treated with thyroidectomy or radioactive ablation of the thyroid gland as well as thyroid-stimulating hormone–suppression therapy. This may lead to disturbances in thyroid hormone levels such as hypothyroidism, symptoms of which include fatigue, poor memory and concentration, weight gain and/or poor appetite, as well as slow pulse and movements . All of these symptoms resemble clinical criteria of depression and are directly assessed by individual items of the PHQ-9. In some patients with this type of cancer, it is thus possible that depressive symptoms may be managed better with adjusting the dose of thyroid hormone replacement than with anti-depressants . Another mechanism may explain the high prevalence of depressive symptoms in pancreatic cancer patients. It has been found that pancreatic tumour cells inhibit t-cell activation in the cancer microenvironment by tryptophan depletion through catabolic enzymes such as indoleamine-2,3-dioxygenase . Based on this observation, it has been hypothesised that the accumulation of neurotoxic tryptophan catabolites and impaired serotonin synthesis are directly responsible for depressive symptoms in these patients such that elevated depressive symptoms could be considered a paraneoplastic syndrome , , and . Indeed an association between plasma kynurenic acid: tryptophan ratio and depressive symptoms (Beck Depression Inventory, BDI) has been observed in a pilot study with 17 pancreatic cancer patients . Similarly, one may observe depressive symptoms in melanoma patients as iatrogenic consequences of adjuvant interferon therapy, potentially mediated by the above-mentioned altered tryptophan catabolism  and .
4.3. Strengths, limitations and conclusions
Strengths of our study include the multi-center design, the large representative sample of cancer patients, including patients with rare cancer types, access to patients' medical records, access to a population-based comparison group and the ability to conduct meaningful non-responder analyses.
The generalisability of our findings may be questioned by several factors. First, we used a cross-sectional design, which prevents us from drawing conclusions on cause and effect. Second, the non-responder analyses revealed a bias towards younger age and higher education. However, the magnitude of these differences was relatively small and we found no evidence of gender bias; we therefore believe the effects to be almost negligible. Despite our efforts of representative sampling, randomisation and self-selection led to an over-representation of patients from cancer rehabilitation centres. As this group was slightly younger, which has been shown to be associated with higher depressive symptoms in this population, it is possible that this led to a small bias towards higher depressive symptoms in our sample. We do not believe this effect to be large and therefore do not see it as a threat to the generalisability of our findings. Third, our results on thyroid, testicular and soft tissue cancer patients should be interpreted with caution, as the samples from these populations were relatively small (between 25 and 39 cases each). Yet, it may also be regarded as a strength that we were able to collect data from these small populations at all. Fourth, we were not able to control general population analyses for somatic co-morbidity or education. Fifth, as our target population consisted of all cancer patients, some patients had a second malignancy, and we were not able to control for prior depression, and genetic factors.
Our data provide health care professionals with reliable estimates on the prevalence of depressive symptoms in cancer patients. Adequately addressing these symptoms does not only improve patients' quality of life but may even increase survival rates by improving treatment adherence and health behaviour. It is therefore essential that health care professionals pay special attention to depressive symptoms in cancer patients and carefully consider relevant treatment.
Conflict of interest statement
This work was supported by the German Cancer Aid (grant number 107465) within the psychosocial oncology funding priority program.
The authors thank all participating patients and staff at participating study centers who helped in data collection.
Appendix A. Supplementary data
The following are the supplementary data related to this article:
-  A.J. Mitchell, M. Chan, H. Bhatti, M. Halton, L. Grassi, C. Johansen, et al. Prevalence of depression, anxiety, and adjustment disorder in oncological, haematological, and palliative-care settings: a meta-analysis of 94 interview-based studies. Lancet Oncol. 2011;12(2):160-174 Crossref
-  J. Walker, C. Holm Hansen, P. Martin, A. Sawhney, P. Thekkumpurath, C. Beale, et al. Prevalence of depression in adults with cancer: a systematic review. Ann Oncol. 2013;24(4):895-900 Crossref
-  M.R. DiMatteo, K.B. Haskard-Zolnierek. Impact of depression on treatment adherence and survival from cancer. D.W. Kissane, M. Maj, N. Sartorius (Eds.) Depression and cancer (Wiley, Chichester, West Sussex, UK, 2011) 101-124
-  A. Mehnert, E. Brähler, H. Faller, M. Härter, M. Keller, H. Schulz, et al. Four-week prevalence of mental disorders in patients with cancer across major tumor entities. J Clin Oncol. 2014;32(31):3540-3546 Crossref
-  S.O. Dalton, T.M. Laursen, L. Ross, P.B. Mortensen, C. Johansen. Risk for hospitalization with depression after a cancer diagnosis: a nationwide, population-based study of cancer patients in Denmark from 1973 to 2003. J Clin Oncol. 2009;27(9):1440-1445 Crossref
-  N.P. Suppli, C. Johansen, J. Christensen, L.V. Kessing, N. Kroman, S.O. Dalton. Increased risk for depression after breast cancer: a nationwide population-based cohort study of associated factors in Denmark, 1998–2011. J Clin Oncol. 2014;32(34):3831-3839 Crossref
-  M. Polidoro Lima, F.L. Osório. Indicators of psychiatric disorders in different oncology specialties: a prevalence study. J Oncol. 2014;2014(32):1-7 Crossref
-  M.J. Massie. Prevalence of depression in patients with cancer. J Natl Cancer Inst Monogr. 2004;32:57-71 Crossref
-  A.M.H. Krebber, L.M. Buffart, G. Kleijn, I.C. Riepma, Bree R. de, C.R. Leemans, et al. Prevalence of depression in cancer patients: a meta-analysis of diagnostic interviews and self-report instruments. Psycho-Oncology. 2014;23(2):121-130 Crossref
-  A. Mehnert, U. Koch, H. Schulz, K. Wegscheider, J. Weis, H. Faller, et al. Prevalence of mental disorders, psychosocial distress and need for psychosocial support in cancer patients – study protocol of an epidemiological multi-center study. BMC Psychiatry. 2012;12:70 Crossref
-  Robert Koch-Institut (Ed.) Krebs in Deutschland 2011/2012. Gesundheitsberichterstattung des Bundes. 10 Ausg. 2015 10th ed. (RKI-Bib1 (Robert Koch-Institut), Berlin, 2015)
-  Deutscher Bundestag. Bundesdatenschutzsgesetz: BDSG.
-  R. Kocalevent, A. Hinz, E. Brähler. Standardization of the depression screener patient health questionnaire (PHQ-9) in the general population. Gen Hosp Psychiatry. 2013;35(5):551-555 Crossref
-  World Health Organization. ICD-10: international statistical classification of diseases and related health problems. 10th ed. (World Health Organization, Geneva, 2011)
-  L.H. Sobin, M.K. Gospodarowicz, C. Wittekind. TNM classification of malignant tumours. 7th ed. (Wiley-Blackwell, Chichester, West Sussex, UK, Hoboken, NJ, 2009)
-  K. Kroenke, R.L. Spitzer, J.B. Williams. The PHQ-9: validity of a brief depression severity measure. J Gen Intern Med. 2001;16(9):606-613 Crossref
-  Löwe B, Spitzer R, Zipfel S, Herzog W. Gesundheitsfragebogen für Patienten (PHQ-D), Zweite Auflage: [PRIME MD Patient Health Questionnaire (PHQ)—German version, 2nd Edition]. Karlsruhe: Pfizer; 2002.
-  B. Löwe, K. Kroenke, W. Herzog, K. Gräfe. Measuring depression outcome with a brief self-report instrument: sensitivity to change of the Patient Health Questionnaire (PHQ-9). J Affect Disord. 2004;81(1):61-66
-  S. Gilbody, D. Richards, S. Brealey, C. Hewitt. Screening for depression in medical settings with the patient health questionnaire (PHQ): a diagnostic meta-analysis. J Gen Intern Med. 2007;22(11):1596-1602 Crossref
-  M.M. Oken, R.H. Creech, D.C. Tormey, J. Horton, T.E. Davis, E.T. McFadden, et al. Toxicity and response criteria of the Eastern Cooperative Oncology Group. Am J Clin Oncol. 1982;5(6):649-655
-  K. Miller, M.J. Massie. Depressive disorders. J. Holland, W. Breitbart, P. Jacobsen, M. Lederberg, M. Loscalzo, R. McCorkle (Eds.) Psycho-oncology 2nd ed. (Oxford University Press, New York, 2010) 311-318 Crossref
-  S. Vehling, C. Lehmann, K. Oechsle, C. Bokemeyer, A. Krull, U. Koch, et al. Global meaning and meaning-related life attitudes: exploring their role in predicting depression, anxiety, and demoralization in cancer patients. Support Care Cancer. 2011;19(4):513-520 Crossref
-  D.M. Clarke. Psychological adaptation, demoralisation, and depression in people with cancer. D.W. Kissane, M. Maj, N. Sartorius (Eds.) Depression and cancer (Wiley, Chichester, West Sussex, UK, 2011)
-  D.L. Longo, A.S. Fauci, D.L. Kasper, S.L. Hauser, J.L. Jameson, J. Loscalzo. 341: disorders of the thyroid gland. D.L. Longo (Ed.) Harrison's principles of internal medicine 18th ed. (McGraw-Hill, New York, 2012)
-  D. Newport, C.B. Nemeroff. Assessment and treatment of depression in the cancer patient. J Psychosom Res. 1998;45(3):215-237 Crossref
-  C. Uyttenhove, L. Pilotte, I. Théate, V. Stroobant, D. Colau, N. Parmentier, et al. Evidence for a tumoral immune resistance mechanism based on tryptophan degradation by indoleamine 2,3-dioxygenase. Nat Med. 2003;9(10):1269-1274 Crossref
-  M. Maes, B.E. Leonard, A.M. Myint, M. Kubera, R. Verkerk. The new ‘5-HT’ hypothesis of depression: cell-mediated immune activation induces indoleamine 2,3-dioxygenase, which leads to lower plasma tryptophan and an increased synthesis of detrimental tryptophan catabolites (TRYCATs), both of which contribute to the onset of depression. Prog Neuropsychopharmacol Biol Psychiatry. 2011;35(3):702-721 Crossref
-  B.M. Campbell, E. Charych, A.W. Lee, T. Möller. Kynurenines in CNS disease: regulation by inflammatory cytokines. Front Neurosci. 2014;8
-  K. Kurz, S. Schroecksnadel, G. Weiss, D. Fuchs. Association between increased tryptophan degradation and depression in cancer patients. Curr Opin Clin Nutr Metab Care. 2011;14(1):49-56 Crossref
-  I.C. Botwinick, L. Pursell, G. Yu, T. Cooper, J.J. Mann, J.A. Chabot. A biological basis for depression in pancreatic cancer. HPB. 2014;16(8):740-743 Crossref
-  P.C. Trask, A.G. Paterson, P. Esper, J. Pau, B. Redman. Longitudinal course of depression, fatigue, and quality of life in patients with high risk melanoma receiving adjuvant interferon. Psycho-Oncology. 2004;13(8):526-536 Crossref
-  D.L. Musselman, A.H. Miller, E.B. Royster, M.D. McNutt. Biology of depression and cytokines in depression. D.W. Kissane, M. Maj, N. Sartorius (Eds.) Depression and cancer (Wiley, Chichester, West Sussex, UK, 2011) 51-80
a Department of Medical Psychology and Medical Sociology, Section of Psychosocial Oncology, University Medical Center Leipzig, Leipzig, Germany
b Department of Psychosomatic Medicine and Psychotherapy, Universal Medical Center Mainz, Mainz, Germany
c Department of Medical Psychology and Psychotherapy, Medical Sociology and Rehabilitation Sciences, and Comprehensive Cancer Center Mainfranken, University of Würzburg, Würzburg, Germany
d Department and Outpatient Clinic of Medical Psychology, University Medical Center Hamburg-Eppendorf, Hamburg, Germany
e Oncology Clinic, 5073 Rigshospitalet, University of Copenhagen, Denmark
f Unit of Survivorship, The Danish Cancer Society Research Center, Copenhagen, Denmark
g Division of Psychooncology, Department for Psychosomatic and General Clinical Medicine, University Hospital Heidelberg, Heidelberg, Germany
h Deanery, University Medical Center Hamburg-Eppendorf, Hamburg, Germany
i Department of Psychooncology, UKF Reha gGmbh University Clinic Center Freiburg, Freiburg, Germany
∗ Corresponding author: Department of Medical Psychology and Medical Sociology, University Medical Center Leipzig, Philipp-Rosenthal-Strasse 55, 04103, Leipzig, Germany. Fax: +49 341 97 15419.
© 2016 Elsevier Ltd, All rights reserved.