Preoperative somatic symptoms are associated with disease progression in patients with bladder carcinoma after cystectomy
A link between patient psychologic factors and bladder carcinoma outcome has not been demonstrated. The purpose of the current study was to assess the association of psychologic factors measured preoperatively with bladder carcinoma progression after cystectomy.
The Brief Symptom Inventory (BSI)-18 was administered prospectively to 65 patients with clinically localized bladder carcinoma before surgery. The BSI-18 measures distress in three specific domains—depression, anxiety, and somatization (i.e., distress due to somatic symptoms)—as well as general distress. Preoperative BSI-18 scores, tumor pathologic stage, and certain clinical variables were compared with disease status. Disease progression was defined as the development of either local disease recurrence or distant metastasis. Univariate and multivariate Cox proportional hazards models were constructed for statistical analysis.
Of the 65 patients, 49 (79.4%) had no evidence of disease, 4 (6.2%) had local disease recurrence, and 12 (18.5%) had metastatic disease at last follow-up. The mean follow-up time was 1.3 years and did not differ significantly between survival outcomes (P = 0.577). Both tumor pathologic stage and preoperative somatic distress scores were associated with time to disease progression by univariate analysis (P = 0.038 and P = 0.055, respectively). After adjusting for tumor pathologic stage, a somatic distress score of ≥ 2.00 was a significant predictor of disease progression (P = 0.044, hazard ratio = 3.31, 95% confidence interval = 1.03–10.60). Patient age, gender, reconstruction type, and BSI-18 scores for depression, anxiety, and general distress were not significantly associated with disease outcome.
The authors found no correlation between psychologic symptoms measured preoperatively (i.e., depression, anxiety, and general distress) and bladder carcinoma progression. However, they reported an association between somatic symptoms and cancer outcome. If confirmed by other studies, these results may have important implications for the diagnosis, staging, and potential treatment of patients with bladder carcinoma. Cancer 2004. © 2004 American Cancer Society.
Many physicians believe that psychologic factors can affect clinical outcome. It has been proposed, for example, that phenomena such as “fighting spirit” and depression have the potential to influence the clinical course of cancer, for better or worse.1 The literature regarding these assumed phenomena, however, is far from conclusive. Although several reports have found a significant association between such patient psychologic factors as fighting spirit, helplessness/hopelessness, and denial with cancer outcome, many others have challenged this assertion.2–4 Even so, the suggestion that psychology can affect biology is not at all unreasonable. Studies of patients with breast carcinoma have revealed increased natural killer cell function and higher cortisol levels in women with perceived social support in comparison to others without the same support network.5, 6 Other reports also have demonstrated altered immune system function and endocrine dysregulation in patients with psychologic stress.7, 8 Unlike psychologic factors, the role of the immune system and neuroendocrine axis in cancer biology is well established. Could psychologic factors, through either a known or unknown mechanism, significantly modulate disease progression in patients with cancer?
Although a significant amount of attention has been paid to this topic in other areas of oncology, relatively little is known about this subject with regards to genitourinary malignancies. The purpose of our study was to assess the influence of psychologic factors, as assessed by the Brief Symptom Inventory- 18 (BSI-18), on disease progression in patients with bladder carcinoma after cystectomy.
MATERIALS AND METHODS
Study Design and Sample Population
The current study was performed prospectively. All patients were diagnosed with clinically organ-confined, either muscle-invasive or carcinoma in situ, transitional cell bladder carcinoma and were seen and evaluated in the Adult Urology Clinic at The Johns Hopkins Hospital (Baltimore, MD). Institutional review board approval and individual patient consent were obtained in the manner set forth in the Johns Hopkins Hospital guidelines for human subjects' research. Individuals with a previous history of a major mental illness were excluded. Patients were asked to complete the BSI-18 during their preoperative visit. Scoring of the BSI-18 was performed by one of the study nurse coordinators. Patients with scores consistent with high levels of psychologic distress were referred for appropriate psychologic consultation. The study population consisted of 74 consecutive patients. Nine patients either did not elect to undergo cystectomy as primary treatment or chose to have surgery elsewhere. Therefore, the study population consisted of 65 patients who completed the preoperative BSI-18 assessment and underwent cystectomy at our institution. Follow-up imaging and diagnostic tests were ordered per the routine at our hospital. The primary end point of disease progression was defined as the development of either local disease recurrence or distant metastasis. Data regarding disease progression were collected in a blinded fashion with regards to BSI-18 scores.
Psychologic Distress Assessment
The BSI-18 was developed as a screening tool to assess psychologic distress in patients with cancer. Its use in the cancer setting has been validated and is well documented.9, 10 The BSI-18 measures distress in three domains: depression, anxiety, and somatization (i.e., distress due to somatic symptoms). The somatic symptoms addressed by the BSI-18 include headaches, dizziness, shortness of breath, fatigue, muscle soreness, and lower back pain. Many of these somatic symptoms can occur as autonomically mediated manifestations of anxiety. After each of the 18 formulated statements (i.e., problem descriptions) included in the questionnaire, patients are asked to specify how much each particular problem has distressed or bothered them (i.e., not at all, a little bit, moderately, quite a bit, or extremely) in the past 7 days. Point values for each response range from 0 for not at all to 4 for extremely. The three distress domains (i.e., depression, anxiety, and somatization) comprise six statements each. The maximum score, therefore, for each of these domains is 24. A general distress score is obtained by summing the scores from the three other domains (i.e., depression, anxiety, and somatization). The maximum score for this scale is thus 72 (i.e., 3 × 24). (A complete description of the BSI-18 may be obtained from the publisher, Pearson Assessments, Inc., Bloomington, MN.) Questionnaires were administered by a study nurse coordinator who used a standardized protocol.
All three distress domains and the general distress score, patient age, pathologic tumor stage, marital status, and urinary reconstruction type were examined for an association with disease progression. Due to the number of patients in the study sample, patients with either local disease recurrence or distant metastasis were grouped for statistical analysis. Initially, univariate proportional hazards analyses were performed to identify potential prognostic factors. Factors associated with cancer progression at P ≤ 0.15 were evaluated in a multivariable Cox proportional hazards model. Somatic distress was a dichotomous variable, rather than a continuous variable, as the distribution of this factor displayed a skewed distribution (i.e., 48% of patients had values equal to 0 preoperatively). Because of the exploratory nature of our study, we tested various cutoff points for somatization and selected the one with the strongest risk estimate. The Student t test was used to compare potential confounders (age, calculated body mass index [BMI], and preoperative hematocrit levels) between patients who had evidence of disease progression and those who did not.
The mean and median follow-up period for the entire patient population was 1.3 years (range, 2.9–0.8 years; Table 1). This did not significantly differ among patients with varying states of disease progression (P = 0.577). The following urinary reconstruction types were performed: 34 ileal conduits, 4 continent pouches, 25 orthotopic neobladders, and 2 ureterostomies. Univariate analysis found that tumor pathologic stage (hazard ratio [HR], 3.23; 95% confidence interval [CI], 1.07–9.72; P = 0.038) and preoperative somatic distress score (HR, 2.89; 95% CI, 0.98–8.52; P = 0.055) were associated with disease progression. Preoperative depression, anxiety, and general distress scores were not significantly associated with cancer outcome. Moreover, age, gender, marital status, and urinary reconstruction type were not significantly associated with disease progression. Multivariable analysis revealed that both tumor pathologic stage and preoperative somatization scores were statistically significant independent predictors of disease progression (Table 2). Indeed, patients with a poorer cancer outcome (i.e., local disease recurrence or distant metastasis) had significantly higher mean preoperative somatic distress scores compared with patients without local disease recurrence or distant metastasis (no evidence of disease, 1.51; local disease recurrence, 3.00; distant metastasis, 4.00; P = 0.012). Despite a significantly higher mean age for patients with somatic distress scores ≥ 2 (65.2 years, compared with 59.8 years for all other patients; P = 0.045), adjustment for age did not significantly influence any of these results. The mean duration of follow-up was not sufficient for an analysis with respect to overall or disease-specific survival.
Table 1. Study Sample Characteristics of Patients with Cystectomy-Treated Bladder Carcinoma Treated Who Completed the Preoperative BSI-18 Assessment
|Mean age (yrs)||62.5||64.5||0.52|
|Male gender (%)||43 (87.8)||12 (75.0)||0.25|
|Pathologic stage|| || || |
| p0||2||0|| |
| pTa||2||0|| |
| pTis||11||1|| |
| pT1||2||1|| |
| pT2||18||3|| |
| pT3||11||10|| |
| pT4||3||1|| |
|Mean BSI-18 scale score (standard deviation)|| || || |
| Depression|| 3.6 (3.9)|| 4.8 (5.3)||0.32|
| Anxiety|| 5.6 (5.6)|| 6.3 (6.6)||0.67|
| Somatization|| 1.5 (2.1)|| 3.8 (3.6)||0.03|
| General Distress||10.9 (10.4)||14.5 (13.7)||0.69|
Table 2. Cox Multivariate Proportional Hazards Model with Regard to Bladder Carcinoma Progression after Cystectomya
|Pathologic stage (T3–4 vs. ≤ T2)||3.06 (1.02–9.20)||0.046|
|Preoperative somatization (score ≥ 2 vs. < 2)||3.31 (1.03–10.60)||0.044|
In accord with many studies in the literature, we found no association between the psychologic phenomena assessed and cancer outcome.4 Both the anxiety and depression scales, as well as the general distress scale, failed to demonstrate a significant correlation with bladder carcinoma recurrence or metastasis after surgery. We did find, however, an association between somatic distress and bladder carcinoma progression after cystectomy. The questions comprising the somatic distress portion of the BSI-18 were formulated to address physical symptoms (e.g., dizziness, shortness of breath) sometimes experienced in anxiety. If one assumes that the somatic symptoms addressed in the BSI-18 are manifestations of anxiety or depression, how can we explain a lack of association between the anxiety and depression scales and bladder carcinoma progression? The simplest explanation is that, in our study sample, somatic symptoms were not a manifestation of anxiety but rather the result of some other process. Two possibilities for this are 1) the somatic symptoms measured in our study were not due to the cancer but to another factor that independently influenced cancer progression, or 2) the somatic symptoms were caused by the cancer itself and herald a more aggressive phenotype.
To be sure, the BSI-18 was developed as a screening measure of psychologic distress, not as a tool to gauge physical symptoms. The finding that patients reported distress attributed to somatic symptoms suggets that the bodily symptoms exist. The severity and/or underlying cause of these symptoms, however, cannot be discerned from our data as our study was not designed to assess this outcome. Patient factors such as obesity and performance status have been shown to modulate cancer outcome in other organ sites independent of cancer stage.11, 12 One could easily see how these same factors might also have an impact on somatic symptoms. A comparison of BMI between patients who experienced cancer progression in our study and those who did not failed to reveal an association (P = 0.174). Furthermore, a review of preoperative hematocrit values, an indirect measure of performance status, also did not reveal a significant difference between the two survival groups (P = 0.98).13 An additional confounding issue may have been the high proportion of male patients (83%) in our study. From a psychologic standpoint, men and women are known to have different ways of expressing psychosocial distress, even in the cancer setting.14 This disparity, however, in and of itself, is unable to explain our results as previous studies have shown that women have a higher prevalence of somatic distress than men and, importantly, as adjustment for gender in our study did not change the multivariable model results.14
What about the possibility that the association found was due to the bladder carcinoma itself? At first glance, one might presume that our results can be explained by the assumption that patients with higher stage tumors and, consequently, a higher incidence of disease progression, have more somatic symptoms. It should be kept in mind that the physical symptoms assessed by the BSI-18 are not those one would typically expect in association with locally advanced bladder carcinoma (i.e., hematuria, dysuria, pelvic pain). Furthermore, in a multivariable analysis accounting for tumor pathologic stage, somatic distress was still significantly associated with bladder carcinoma progression. A more provocative explanation for our findings involves the production of humoral factors either by the tumor or in response to it, which could potentially influence tumor biology as well as patient distress. Studies in solid tumor oncology, including studies of the genitourinary system, have revealed elevated serum cytokine levels, especially interleukin (IL)-6, IL-8, and IL-10, in patients with a variety of high-grade cancers.15–17 Elevated serum IL-6 levels have been associated with symptoms of depression in patients with cancer and also have been shown to modulate the hypothalamic-pituitary-adrenal (HPA) axis.18, 19 Perturbations of the HPA axis are capable of affecting patient psychologic and somatic symptoms. In addition, there are reports that serum IL-6 levels correlate with pain perception postoperatively.20 But, perhaps, the most compelling evidence in support of this hypothesis is that high serum IL-6 activity has been detected in patients with bladder carcinoma and has been significantly associated with disease recurrence and survival after cystectomy.21 These data provide an intriguing possible explanation for our findings. Although paraneoplastic syndromes associated with transitional cell carcinoma of the bladder are extremely rare, none reported thus far are capable of accounting for our findings.22
We cannot exclude the possibility that our results are due to chance or confounding that could not be controlled with the data available. The binary cutoff point used as an indicator of somatic distress (score ≥ 2 vs. ≤ 1) was relatively low. No other binary cutoff point achieved a statistically significant effect, although all demonstrated HRs > 2, and the use of a continuous variable yielded results that approached statistical significance (P = 0.064). Future studies involving a larger sample size and longer follow-up, with a combined approach consisting of appropriate medical and surgical evaluation, suitable psychologic assessment, and examination of serum specimens for specific cytokines, such as IL-6, are needed to corroborate and more fully characterize our findings.
In conclusion, in the current study, we found preoperatively assessed somatic symptoms to be significantly associated with disease progression in patients with bladder carcinoma treated with cystectomy. No such correlation was seen with regards to depression, anxiety, and general distress scales. Our findings, if replicated and elaborated, may have important implications for the diagnosis, staging, and treatment of patients with bladder carcinoma.