Background: Depression is associated with an increased risk of mortality in patients with cancer; it has been hypothesized that depression-associated alterations in cell aging mechanisms, in particular, the telomere/telomerase maintenance system, may underlie this increased risk. We evaluated the association of depressive symptoms and telomere length to mortality and recurrence/progression in 464 patients with bladder cancer.
Methods: We used the Center for Epidemiologic Studies Depression Scale (CES-D) and Structured Clinical Interview for DSM-IV Disorder (SCID) to assess current depressive symptoms and lifetime major depressive disorder (MDD), respectively, and telomere length was assessed from peripheral blood lymphocytes. Multivariate Cox regression was used to assess the association of depression and telomere length to outcomes and the joint effect of both. Kaplan–Meier plots and log-rank tests were used to compare survival time of subgroups by depression variables and telomere length.
Results: Patients with depressive symptoms (CES-D ≥ 16) had a 1.83-fold [95% confidence interval (CI), 1.08–3.08; P = 0.024] increased risk of mortality compared with patients without depressive symptoms (CES-D < 16) and shorter disease-free survival time (P = 0.004). Patients with both depressive symptoms and lifetime history of MDD were at 4.88-fold (95% CI, 1.40–16.99; P = 0.013) increased risk compared with patients with neither condition. Compared to patients without depressive symptoms and long telomere length, patients with depressive symptoms and short telomeres exhibited a 4-fold increased risk of mortality (HR, 3.96; 95% CI, 1.86–8.41; P = 0.0003) and significantly shorter disease-free survival time (P < 0.001).
Conclusion: Short telomere length and depressive symptoms are associated with bladder cancer mortality individually and jointly.
Impact: Further investigation of interventions that impact depression and telomere length may be warranted in patients with cancer. Cancer Epidemiol Biomarkers Prev; 24(2); 336–43. ©2014 AACR.
This article is featured in Highlights of This Issue, p. 321
In the United States, bladder cancer is the fourth most frequently diagnosed cancer in men and eleventh in women, with an estimated 55,600 new cases and 10,510 deaths in 2012 (1). During the past 20 years, treatment of bladder cancer has not changed significantly and the outcomes for patients remain suboptimal (2). Even though grade, stage, and treatment are important prognostic predictors of bladder cancer outcome, current prediction models and nomograms are limited in accuracy and lack external validation (3). Incorporation of novel prognostic variables into current models and simultaneous evaluation of multiple factors has potential to improve outcome prediction (4).
There is compelling evidence that depression is associated with elevated mortality in the general population (5), and poor survival in aging-related diseases (6, 7). Prospective studies have found depression to be associated with 20% to 40% increased mortality in patients with cancer (8, 9), although the association between depression and cancer incidence, recurrence, and progression remains inconclusive (9). Most studies have included mixed cancer types, or evaluated outcomes of common cancers, such as breast cancer (10, 11). No study has been conducted to assess the impact of depression on mortality in patients with bladder cancer.
It has been hypothesized that depression may affect cancer mortality and progression through shared biochemical pathways that lead to an increase in cell damaging processes and a decrease in cell protective or restorative processes (12). One way in which such processes may mediate the effects of depression on cancer outcomes is through alterations in cell aging mechanisms, in particular, the telomere/telomerase maintenance system (12).
Telomeres are specialized repetitive DNA sequences at the ends of chromosomes that protect chromosomes and are critical in maintaining genomic integrity (13). Telomere attrition results in cellular senescence and apoptosis (13). Telomere dysfunction has been linked to many aging-related human diseases, including cancer. Several studies have found a relationship between shorter telomere length and major depressive disorder (MDD; refs. 14–16).
Compared with the extensive studies of telomere length in the context of modifiable risk factors (17) and disease risks (18), research on the relationship between telomere length and disease prognosis is sparse. One series of studies that prospectively examined the association between telomere length and mortality in patients with cancer (19, 20) found that shorter telomere length was associated with increased all-cause mortality of patients with cancer over a 10- and 15-year period.
The primary objective of the current study was to prospectively evaluate the association between baseline depressive symptoms, symptoms of lifetime MDD, and telomere length to all-cause mortality, recurrence, and disease progression in 464 patients with bladder cancer. We hypothesized that depressive symptoms, history of MDD, and short telomere length would each be associated with increased mortality and recurrence/progression, and that patients with current depressive symptoms and history of MDD would have shorter survival and greater likelihood for recurrence/progression than those who met criteria on only one of the depression measures. A secondary objective was to evaluate whether telomere length moderated the relationship between depressive symptoms/lifetime MDD and mortality and recurrence/progression outcomes, such that those with depressive symptoms and history of MDD with short telomere length would have the shortest survival time and greater likelihood for progression. We also evaluated whether telomere length mediated the association between depression symptoms and history of MDD and mortality and recurrence/progression.
Materials and Methods
Study population, data collection, and laboratory assays
All of the human participation procedures were approved by the University of Texas MD Anderson Cancer Center Institutional Review Boards. Written informed consent was obtained from each participant before enrollment.
Patients with bladder cancer were recruited in an ongoing epidemiologic study of bladder cancer at The University of Texas MD Anderson Cancer Center and Baylor College of Medicine (Houston, TX). Patient recruitment began in 1999. Procedures for recruitment and eligibility criteria have been described previously (21). Briefly, incident bladder cancer patients were identified through a daily review of computerized appointment schedules. Patients were diagnosed within one year upon recruitment, were histologically confirmed with urinary bladder cancer, and had not received prior chemotherapy or radiotherapy before enrollment. There were no recruitment restrictions on age, gender, ethnicity, or stage.
After written informed consent was obtained, trained MD Anderson staff interviewers administered a 45-minute structured risk factor questionnaire to collect data on demographic characteristics, socioeconomic status, detailed tobacco use history, lifestyles, occupational history, medical history and medications, and family history of cancer.
Current level of depressive symptom severity was assessed using the Center for Epidemiologic Studies Depression Scale (CES-D), a 20-item self-report measure developed to assess depressive symptoms in community populations (22). Patients with a CES-D score ≥ 16 were classified at screening as meeting criteria suggestive of current depression (23).
Lifetime history of MDD was measured using the two cardinal items for major depressive episode from the Structured Clinical Interview for DSM-IV Disorder (SCID; ref. 24); these items assessed lifetime presence of depressed mood or anhedonia for 2 weeks or more. Patients were considered to have lifetime MDD if they endorsed one or both of the items.
Clinical data, including tumor size, grade, stage, presence of carcinoma in situ (CIS), number of tumor foci at diagnosis, intravesical therapy, dates of recurrence and progression events, systemic chemotherapy, radical cystectomy, pathologic findings at cystectomy, and vital status, were abstracted from medical charts. The date of death or date of last follow-up for patients alive or lost to follow up was recorded. The MD Anderson Tumor Registry conducts annual vital status follow-ups on all patients with cancer. Computer matches are performed with MD Anderson appointment files to identify patients who have presented for a recent appointment. For those patients who have not had a visit within the past 12 to 15 months, letters are sent to the patient to ask about the patient's health and vital status. If the patient does not respond to the letter, phone calls are made, and the social security death index is checked. The endpoint outcomes in this study included all-cause mortality, which was calculated from the date of diagnosis to the date of death or last follow-up, whichever came first. In addition, recurrence and progression were also recorded for non–muscle-invasive bladder cancer (NMIBC). Recurrence was defined as a newly found bladder tumor following a previous negative follow-up cystoscopy, and progression was defined as the transition from non–muscle-invasive to invasive or metastatic tumors (25).
Immediately after the interview, a 40 mL blood sample was collected in sodium-heparinized tubes. Genomic DNA was isolated from peripheral blood using the QIAamp DNA Blood Maxi Kit (Qiagen) according to the manufacturer's protocol. The DNAs were stored in −80°C freezer for future use.
The relative overall telomere length was measured by real-time quantitative PCR as previously described (26). Briefly, the relative overall telomere length is proportional to the ratio (T/S) of the telomere repeat copy number and the single (human globulin) copy number. The T/S ratio of each sample was normalized to a calibrator DNA to standardize between different runs. The PCR reaction mixture (15 μL) for telomere amplification consists of 5 ng genomic DNA, 1 × SYBR Green Mastermix (Applied Biosystems), 200 nmol/L Tel-1 primer (CGGTTTGTTTGGGTTTGGGTTTGGGTTTGG-GTTTGGGTT), and 200 nmol/L Tel-2 primer (GGCTTGCCTTACCCTTACCCTTACCCTTACCCTTACCC-T). The PCR reaction mixture (15 μL) for human globulin consists of 5 ng genomic DNA, 1 × SYBR Green Mastermix, 200 nmol/L Hgb-1 primer (GCTTCTGACACAACTGTGTTCACTAGC), and 200 nmol/L Hgb-2 primer (CACCAACTTCATCCACG-TTCACC). The thermal reactions are: 95°C for 10 minutes followed by 40 cycles of 95°C for 15 seconds and 56°C (for telomere amplification) or 58°C (for Hgb amplification) for 1 minute. The PCR reaction for telomere and Hgb amplification were done in two separate 384 plates with the same arrangement of samples on plates. Each plate, besides samples, included negative and positive controls (3.9 Kb telomere and 1.2 Kb telomere from Roche Applied Science commercial telomere length assay kit), and a calibrator DNA. Each plate contained six 2-fold serially diluted reference DNA (from 20 to 0.625 ng) for a six-point standard curve. The acceptable SD of Ct values was ≤0.25. Otherwise, the sample was repeated. The interassay variations were tested by two samples for three runs.
All statistical analyses were performed using the Intercooled Stata 10 statistical software package (Stata) and were two-sided. P <0.05 was considered as statistically significant. To evaluate the effect of depressive symptoms, lifetime MDD, and telomere length on bladder cancer mortality, HRs and 95% confidence intervals (95% CI) were estimated by Cox proportional hazard regression for the overall analysis, and stratified analysis by stage. Patients who were lost to follow-up before the endpoint events (death, recurrence, and progression) were censored. Univariate Cox regression was performed to evaluate the association between outcomes with demographic, clinical variables, telomere length, and depression variables. Variables that were significantly associated with bladder cancer outcomes in univariate Cox regression were included in the multivariate Cox regression model.
The CES-D score was analyzed as both categorical (CES-D scale ≥16 or <16) and continuous. Telomere length was also analyzed both as a continuous and categorical variable by median split and quartiles to assess dose–response trends. Furthermore, in a sensitivity analysis, we used age as the time scale in the Cox proportional hazard regression and observed similar estimates for HRs and 95% CIs for the depressive symptoms, lifetime MDD, and telomere length (data not shown). The Kaplan–Meier plots and log-rank tests were applied to compare the difference in event-free survival time by the dichotomized groups for current depressive symptoms, lifetime MDD, and telomere length, where appropriate. Median survival time of each group was estimated based on the Kaplan–Meier curve. Joint association of short telomere and depression measures was assessed by both multivariate Cox regression and Kaplan–Meier curves. Multivariate logistic regression was used to assess the association between telomere length and depression variables.
Mediation analyses were conducted using standard procedures (27) from multivariate regression analyses noted above. The criterion for mediation requires that the predictor variables (CES-D/MDD) are associated with telomere length and mortality; telomere length (mediator) is associated with mortality; and when both CES-D/MDD and telomere length are entered in the same model, the association between CES-D/MDD and mortality is decreased.
Of the 464 patients with bladder cancer enrolled (Table 1), 368 (79.3%) were male and 432 (93.1%) were Caucasian. The mean age at diagnosis was 65 years. There were 234 (53.7%) patients with NMIBC and 202 (46.3%) with muscle-invasive and metastatic bladder cancer (MIBC). At follow-up, clinical data were available for 441 of the 464 patients enrolled. During the median follow up of 21.6 months, 88 patients died, 124 patients experienced recurrence, and 73 patients progressed from NMIBC to a more advanced stage.
Relationship of depression to mortality.
Age at cancer diagnosis, cancer stage, grade, and treatment for MIBC were all significantly associated with overall survival (Table 2). The presence of current depressive symptoms (CES-D scores ≥ 16) was significantly associated with increased risk of bladder cancer mortality (HR, 1.90; 95% CI, 1.21–2.98; P = 0.005; Table 2). The association between bladder cancer mortality and current depressive symptoms was also confirmed using CES-D as a continuous measure. However, lifetime MDD history alone was not significantly associated with bladder cancer survival (Table 2). In multivariate Cox regression analysis (Table 3), after adjusting for age, gender, ethnicity, smoking status, cancer stage, grade, and treatment regimens, patients with current depressive symptoms exhibited a 1.83-fold (95% CI, 1.08–3.08; P = 0.024) elevated mortality risk, and shorter median survival time, based on Kaplan–Meier survival analysis (58.26 months vs. greater than 200 months; P = 0.004; Supplementary Fig. S1), compared with those with CES-D scores < 16. Patients with lifetime MDD also showed increased mortality in multivariate analyses, although the association did not reach statistical significance (Table 3). In joint multivariate analysis of lifetime MDD and current depressive symptoms, patients with current depressive symptoms and lifetime MDD experienced 4.88-fold (95% CI, 1.40–16.99; P = 0.013) increased risk of mortality compared with those with CES-D scores < 16 and no lifetime MDD (Table 4). Patients with both forms of depression had significantly shorter median survival time (44 months vs. greater than 200 months; P = 0.005; Fig. 1).
Relationship of telomere length to mortality.
In univariate analyses, one unit increase in telomere length was associated with a significantly reduced risk of death (HR, 0.39; 95% CI, 0.20–0.78; P = 0.007; Supplementary Table S1). When dichotomized by median, long telomere length was associated with a 45% reduction in mortality (HR, 0.55; 95% CI, 0.34–0.89; P = 0.015; Supplementary Table S1). Reduced mortality was observed with long telomere length in quartile analyses suggesting a significant dose–response relationship between quartiles and mortality (P for trend = 0.018 in quartile analysis; Supplementary Table S1). However, in multivariate models adjusting for age, gender, ethnicity, smoking status, tumor stage, grade, and treatment, the associations were attenuated and did not reach statistical significance (Supplementary Table S1). Kaplan–Meier survival curves for median split showed that median survival time was significantly shorter in patients with short versus long telomere lengths (58.26 months vs. greater than 200 months; P = 0.013; Fig. 2).
Relationship of telomere length and depression to recurrence/progression.
In the subset of superficial bladder cancer patients, due to the small sample size and endpoint events, we combined recurrence and progression as one endpoint event. The recurrence/progression-free survival times did not differ significantly between subjects with short versus long telomere lengths. However, multivariate Cox regression analysis indicated that patients with a lifetime history of MDD had a 1.99-fold elevated risk for having a recurrence or progression of disease compared with those with no history of MDD (OR, 1.99; 95% CI, 1.14–3.47; P = 0.016; Supplementary Table S2).
Joint effect of depression and telomere length on mortality.
In joint analysis, patients with both short telomere length and CES-D scores ≥ 16 experienced a 3.96-fold (95% CI: 1.86–8.41; P = 0.0003) increased mortality compared to patients with long telomeres and CES-D scores < 16 (Table 4), as well as shorter median survival time as compared with other groups (31 months vs. greater than 200 months, P < 0.001; Fig. 3). A similar joint analysis of MDD history and telomere length on mortality was nonsignificant (Table 4). However, patients with lifetime MDD and short telomere length had significantly shorter median survival time than other groups (58.26 months vs. greater than 200 months; P = 0.025; Supplementary Fig. S2).
Evaluation of telomere length as mediator.
There were no direct associations between CES-D scores and telomere length (Supplementary Table S3). However, in multivariate Cox regression analysis, history of lifetime MDD was associated with increased risk of having short telomere length (OR, 1.85; 95% CI, 1.03–3.32; P = 0.039; Supplementary Table S3). Furthermore, the mean telomere length was significantly shorter in patients with lifetime MDD than patients with no history of MDD (1.23 vs. 1.42, P = 0.002). However, the criterion for meditational analyses did not support the hypothesis of telomere length as a mediator of the effects of CES-D scores or history of MDD on mortality.
We found that current depressive symptoms were associated with increased bladder cancer mortality by almost 2-fold and that current depressive symptoms and lifetime history of MDD further increased the risk to almost five fold. We did not, however, find evidence to support the hypothesis that telomere length mediated these effects. Our results did support a strong joint association of short telomere length and current depressive symptoms with mortality (3.96-fold).
Several hypotheses have been proposed to explain the association between depression and increased mortality risk in patients with cancer (8, 28). The association may be due to symptom overlap between depressive symptoms and disease severity. However, previous meta-analyses have found depressive symptoms to be predictive of cancer mortality, independent of a number of clinical prognosticators including stage of disease, cancer site, performance status, treatment status, and smoking (8, 9), as was the case in the current trial. The association may also be due to inequities in access to care for patients with depression which could result in lower receipt of appropriate cancer treatments following diagnosis (29, 30), although receipt of appropriate treatment was not found to mediate the relationship of psychiatric disorder and mortality risk in these studies (29, 30). Depressed individuals may be more likely to engage in behaviors that negatively impact health, including smoking and drinking alcohol in excess, overeating, engaging in low levels of physical activity, and poor sleep hygiene. In the current study, we found depression was associated with mortality, independent of smoking. Although data on these other behavioral factors was not available in the current study, other studies have shown that the effect of depression on mortality was attenuated after adjusting for these behavioral factors, but remained a significant predictor (31, 32).
Alternatively, depression may enhance cancer mortality via neuroendocrine and immunologic mechanisms (28, 33). MDD is associated with abnormalities in stress-related biologic systems including the sympathetic nervous system, hypothalamic–pituitary–adrenal axis (HPA); and immune function, including increased circulating cortisol concentrations, inflammation, free radical production, and oxidative stress (12, 34), which could lead to immune response suppression to tumors. Increased concentrations of neutrophils and the expression of proinflammatory interleukins such as IL1, IL6, TNFα, and other inflammatory factors (35), as well as an imbalance between Th1 and Th2 immune response have been found in depressed patients (35). Depression is associated with decreased activity of cytotoxic T cells and natural killer cells, which may affect immune surveillance of tumors (28, 35). Dysfunctional immune surveillance could lead to the accumulation of somatic mutations, genome instability, tumorigenesis, and progression of disease (35). Other cancer-related events such as DNA damage, angiogenesis, cell growth, and reactive oxygen species have also been linked to depression (35).
A significant finding of the current study is that patients with both depressive symptoms and short telomere length had a higher level of mortality risk. The neuroendocrine and immunologic processes associated with depression, noted above, have also been found to be related to telomere length, telomerase activity (12), and tumor growth and progression (36). The joint association of depression and telomere length with mortality might suggest an additive effect of these depression-associated processes to other factors we found to be associated with telomere length including age and type of treatment in MIBC (chemotherapy and transurethral resection).
Although we found telomere length was associated with mortality in univariate analyses, the association was nonsignificant in multivariate analyses that adjusted for demographic and clinical variables. Our findings differ from those of multivariate analysis findings in a longitudinal population-based study of cancer-free individuals (19), and studies of recently diagnosed cancer patients (37, 38) that found telomere length was associated with cancer mortality. Discrepancies in findings may be due to differences in telomere length assessment methodology, tumor type, and telomere length response to different types of cancer disease.
Contrary to our hypothesis, we did not find depressive symptoms were associated with telomere length. However, we did find an inverse association between lifetime MDD and telomere length. A majority of studies examining the relationship between depressive episodes in MDD and bipolar disorder found shorter telomeres and a higher load of short telomeres in individuals with these disorders compared with nonpsychiatric controls (14–16, 39), and findings from several studies suggest that depressive chronicity may play an important role in the association between depression and telomere length (14, 16). In contrast, a number of studies have failed to find an association between telomere length and severity of depressive symptoms assessed by self-report measures such as the CES-D (40, 41). Our results are consistent with these findings. It is important to note that studies examining the effects of chronic psychological stress on telomere length have also found a negative dose–response relationship of telomere length with increasing levels of stress and traumatic life events in adults and children (42–44), which could suggest that chronic exposure to the biologic effects of psychological stress (12) may be a common mechanism underlying the association of telomere length and these various forms of stress exposure.
The current study is the first to assess the association between telomere length and bladder cancer outcomes. Results from previous studies examining the association between telomere length and cancer mortality were inconsistent by cancer type, with mortality in some cancers related to short telomere length (45), and long telomere length in other cancers (37, 38). However, results from most of these studies were based on small samples, and/or had limited control for confounding factors, potentially leading to unreliable risk estimates.
Cognitive behavioral, pharmacologic, and combined treatment approaches have been found to be efficacious in the treatment of depressive symptoms, and should be considered in patients with cancer with MDD or those who score high on measures of depressive symptoms. Moreover, telomere shortening has been linked to a number of modifiable environmental and lifestyle factors that contribute to disease risk and progression such as smoking, pollution, obesity (46, 47), and early and chronic life stress (44, 48). Higher engagement in healthy behaviors, including diet, exercise and sleep has been found to moderate, in a protective fashion, the effect of major life stressors on telomere attrition across a 1-year period in healthy, older women. In addition, there is preliminary evidence that interventions designed to impact diet, physical activity, stress, and social support are associated with increases in telomerase activity and telomere length in patients with low-risk prostate cancer (49, 50). Additional research is needed on the effect of behavioral interventions on telomere attrition.
Current bladder cancer outcome prediction has largely relied on clinical parameters (3, 4, 51), which has resulted in low predictive power with appreciably different outcomes in patients with similar clinical characteristics. The current study strongly suggests that incorporating psychological risk factors into outcome models could provide additional and important information regarding predictors of risk. Depressive symptoms remained statistically significant in the model after adjusting for demographic and clinical variables, supporting depression as an independent predictor of mortality in patients with bladder cancer. Because depression has biologic effects, as previously discussed, incorporating depression into current prediction models may account for the effects of biologic predictors that have not yet been identified or characterized.
There were several limitations in this study. Because of the observational nature of the study, a causal link between depression, telomere length, and survival cannot be established. However, the data are consistent with well controlled animal studies examining the role of stress in progression of disease (52). We used a measure of depressive symptoms rather than clinical diagnosis. Our measure of lifetime MDD was based on just two questions, and the measure we used has not yet been validated. However, it should be noted that a two-item screen for current MDD that utilizes the same two cardinal questions, has been found to have high levels of sensitivity and specificity in patients with cancer and other clinical populations (53–56). Furthermore, our measure of lifetime MDD did not provide information on whether participants were experiencing a current major depressive episode, level of symptom severity, or the duration or number of previous major depressive episodes. This is an important limitation, given recent findings from a large, population-based study, showing a dose–response association between severity and chronicity of symptoms and telomere length in individuals with current MDD, with the most severely disordered individuals having the shortest telomeres. However, this study also found individuals with remitted MDD to have significantly shorter telomeres than healthy controls, and no differences in telomere length in individuals with current versus remitted MDD (57). Finally, depression was measured only once at study entry, thus the temporal relationship of depressive symptom changes to cancer outcomes could not be assessed.
This is the first prospective study investigating depressive symptoms, telomere length, and mortality in a cohort of patients with bladder cancer. We observed a joint association of depressive symptoms and short telomere length with all-cause mortality. Our findings have significant public health implications regarding the importance of intervening on depressive symptoms and utilizing behavioral interventions that may slow telomere attrition rate in patients with cancer.
Disclosure of Potential Conflicts of Interest
No potential conflicts of interest were disclosed.
Conception and design: J.A. Blalock, M. Chen, L. Cohen, P.M. Cinciripini, X. Wu
Development of methodology: J.A. Blalock, M. Chen, P.M. Cinciripini, X. Wu
Acquisition of data (provided animals, acquired and managed patients, provided facilities, etc.): M. Chen, P.M. Cinciripini, X. Wu
Analysis and interpretation of data (e.g., statistical analysis, biostatistics, computational analysis): J. Lin, J.A. Blalock, M. Chen, Y. Ye, L. Cohen, P.M. Cinciripini, X. Wu
Writing, review, and/or revision of the manuscript: J. Lin, J.A. Blalock, M. Chen, Y. Ye, J. Gu, L. Cohen, P.M. Cinciripini, X. Wu
Administrative, technical, or material support (i.e., reporting or organizing data, constructing databases): X. Wu
Study supervision: J. Gu, X. Wu
This work was supported by the NIH (U01 CA 127615; to X. Wu) and P50 CA 91846 (to X. Wu and C.P. Dinney).
The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.