Authors: Michael J. Robinson (1Vanderbilt Ingram Cancer Center, Vanderbilt University Medical Center, Nashville, TN), Sang Minh Nguyen (2Division of Epidemiology, Department of Medicine, Vanderbilt Epidemiology Center, Nashville, TN), Debra L. Friedman (1Vanderbilt Ingram Cancer Center, Vanderbilt University Medical Center, Nashville, TN), Emma A. Schremp (1Vanderbilt Ingram Cancer Center, Vanderbilt University Medical Center, Nashville, TN), Lucy L. Wang (1Vanderbilt Ingram Cancer Center, Vanderbilt University Medical Center, Nashville, TN), Scott C. Borinstein (1Vanderbilt Ingram Cancer Center, Vanderbilt University Medical Center, Nashville, TN), Elizabeth J. Davis (1Vanderbilt Ingram Cancer Center, Vanderbilt University Medical Center, Nashville, TN), Tuya Pal (1Vanderbilt Ingram Cancer Center, Vanderbilt University Medical Center, Nashville, TN), Ben H. Park (1Vanderbilt Ingram Cancer Center, Vanderbilt University Medical Center, Nashville, TN), Xiao-Ou Shu (1Vanderbilt Ingram Cancer Center, Vanderbilt University Medical Center, Nashville, TN; 2Division of Epidemiology, Department of Medicine, Vanderbilt Epidemiology Center, Nashville, TN)
Categories: Article
Source: JCO oncology practice
Doi: 10.1200/OP.24.00163
Authors: Michael J. Robinson, Sang Minh Nguyen, Debra L. Friedman, Emma A. Schremp, Lucy L. Wang, Scott C. Borinstein, Elizabeth J. Davis, Tuya Pal, Ben H. Park, Xiao-Ou Shu
Prevalence and risk factors for depression among patients with sarcoma and survivors of sarcoma are not well characterized.
A sarcoma survivorship cohort was constructed from patients diagnosed between April 2022 and September 2023. Depression symptoms were assessed via the eight-item Patient-Reported Outcomes Measurement Information System-57 depression scale at enrollment. Standardized T-score levels (<50, 50–59, and ≥60) were calculated and evaluated in association with demographics, lifestyle characteristics, clinical data, and modifiable factors using multinomial logistic regression models.
Among 612 participants, the mean T-score was 48.3 (standard deviation, 10.0); 58.8% had a T-score <50, 27.9% scored between 50 and 59, and 13.2% scored ≥60. Participants age 18–39 years and age 40–59 years were more likely to have a T-score ≥60, with respective odds ratios (ORs) of 3.65 (95% CIs, 1.70 to 7.83) and 2.80 (1.52 to 5.17) compared with participants older than 60 years. Household incomes of 120,000 in US dollars (USD) (OR, 0.46 [95% CI, 0.23 to 0.92]) and >45,000 USD. Marijuana use within the past 30 days was positively (OR, 3.48 [95% CI, 1.46 to 8.27]) associated, while regular exercise (OR, 0.43 [95% CI, 0.24 to 0.75]) and emotional support (OR, 0.37 [95% CI, 0.28 to 0.48]) were inversely associated with having T-score ≥60.
A higher prevalence of depression symptoms was notable in younger participants, marijuana users, and households with lower incomes. Regular exercise and increased emotional support were inversely associated with depression symptoms. Our study provides information for developing personalized supportive care strategies to ameliorate depression symptoms among patients with sarcoma.
Advancements in sarcoma treatment have led to an overall 5-year survival rate of approximately 65.5%, with variation on the basis of histology, site, tumor size, and stage.^1–7^ However, sarcoma therapy is often associated with adverse physical, psychosocial, and neurocognitive outcomes, affecting the quality of life among survivors.^8–18^
There is indeed a unique association between sarcoma and depression. Previous studies have demonstrated the prevalence of depression among patients with sarcoma to be approximately 35%, nearly twice the national level of 18.4%, as reported by the Centers for Disease Control and Prevention (CDC).^19,20^ This link is likely secondary to the intensity of treatment, as well as social and mobility changes associated with a sarcoma diagnosis.^21–25^ The correlation of patient or survivor demographics, lifestyle, and disease characteristics with symptoms of depression is an area of insufficient and limited quantity of data.
Studies have reported an increased prevalence of depression among patients with cancer as a singular population, not distinguishing between cancer subtypes.^26–28^ Factors most consistently studied include age, race, sex, exercise, marital status, employment status, education, and substance use. However, there are inconsistencies among the results of such studies, signaling that further research is needed, especially among patients with unique diagnoses and needs. Furthermore, many previous studies were limited by sample size, lack of standardized measurements, or performed in countries with dissimilar health care policies compared with the United States.^18,21,29–37^ There are scarce data evaluating potential risk factors for depression among patients with sarcoma, across a wide age spectrum, treated in the United States.
With increasing survivorship, evaluating determinants of depression specific to sarcoma is critical. Clinical depression increases an individual’s risk for multiple adverse outcomes, including poor physical health, treatment noncompliance, substance use, social and functional impairments, diminished quality of life, and suicide.^38–51^ Data suggest that patients with simultaneous diagnoses of cancer and depression (and/or anxiety) have increased health care costs, emergency room visits, and hospitalizations.^19,52,53^ Additionally, the majority of individuals with cancer who experience significant levels of distress do not see mental health professionals.^54^
To research these gaps, we comprehensively evaluated the prevalence of, and risk factors for, depression in a cohort of patients with sarcoma in active therapy or post-therapy surveillance using validated measures. Identifying factors that are associated with depression is a critical first step in establishing proactive measures for at-risk patients and survivors of sarcoma, providing a foundation for the development of supportive care strategies to patients with an increased risk of developing depression.
This study used data from the Cohort to Augment the Understanding of Sarcoma Survivorship Across the Lifespan (CAUSAL) study, an ongoing cohort study designed to assess the impact of sarcoma, its treatment, and patient characteristics on health outcomes. A total of 895 adult patients with sarcoma (age range, 18–86 years) treated at Vanderbilt University Medical Center (VUMC) were enrolled in the study between April 2022 and September 2023. Written consent was obtained from all participants, and approval for human subject research was obtained from VUMC.
At study enrollment, patient surveys were conducted through Research Electronic Data Capture to collect information on patient demographic characteristics and lifestyle factors. Clinical characteristics and sarcoma treatment data (including systemic chemotherapy, radiation, surgery, and targeted therapy) were abstracted from the electronic health records (EHRs).
Self-reported depression symptoms were assessed via the Patient-Reported Outcomes Measurement Information System (PROMIS)-57, which encompasses domains across physical, mental, and social health in adults.^55^ Depression was assessed by the eight-item depression scale, which asked participants how frequently in the past 7 days they experienced various depression symptoms, including feeling worthless, helpless, depressed, hopeless, like a failure, unhappy, having nothing to look forward to, or nothing that could cheer them up. These eight items were scored on a five-point Likert scale with responses of never (1), rarely (2), sometimes (3), often (4), and always (5). A raw score was calculated and then converted into a standardized T-score with a population mean of 50 and a standard deviation (SD) of 10.^55^ Higher T-scores were associated with more symptoms of depression. Depression T-scores were categorized into three T-score <50 (below the mean of the general population; low prevalence of depression symptoms), T-score of 50–59 (within one SD of the mean of the general population), and T-score ≥60 (greater than one SD above the mean of the general population; high prevalence of depression). Participants with incomplete responses were excluded (n = 283), leaving 612 cases included in the current analysis.
Descriptive statistics of patients’ demographic characteristics were computed in percentages for categorical variables, and mean and SD for continuous variables. Differences across groups were compared using the chi-squared test or Wilcoxon rank-sum test. Associations of participant characteristics, clinical factors, and modifiable factors with the depression T-score levels (T-score <50 v T-score 50–59 and T-score ≥60) were evaluated using multinomial logistic regression models. Odds ratios (ORs) and 95% CIs were calculated with adjustments for covariates. Covariates evaluated included age groups at recruitment (18–39, 40–59, and ≥60), sex (male/female), race (White/others), marriage status (married or live with a partner, divorced/windowed/separated, or never married), education (less than college/college or higher), household income (<45,000–70,000–120,000 USD), BMI level (<25, 25–29, or ≥30 kg/m^2^), regular exercise (yes/no), alcohol use (yes/no), smoking (never smoking, ever smoking, or current smoking), marijuana use (yes/no), tumor type (soft tissue sarcoma/bone cancer), tumor site (extremity/others), medication used for depression or pain during 30 days before depression assessment (yes/no), emotional support score (a total score of the 16-item five-point Likert scale; continuous in per-SD increase), and interval from the last previous cancer treatment to depression assessment (eg, no treatment, <1 year, >1 year). No treatment was defined as the participant not having received any previous therapeutic intervention (such as chemotherapy, radiation, surgery, or targeted therapy) before enrollment and completion of the PROMIS measure. Regular exercise was defined by self-reported regular participation in leisure time physical activities within 12 months before enrollment, while alcohol use was obtained on the basis of the self-reported number of days participants drank at least one alcoholic beverage during the previous 30 days. Additionally, associations of participant characteristics, clinical factors, and modifiable factors with the per-SD increase of depression T-score were evaluated using multivariable linear regression analysis with the mean differences (β coefficients) and 95% CIs as effect estimates. A two-sided P value of <.05 was considered statistically significant. All statistical analyses were performed using R version 4.3.2.
The analytic cohort included 612 study participants, with a median of 2.04 years since the initial diagnosis, of whom 59.8% received no previous treatment, while 29.9% and 20.1% had their most recent treatment less than 1 year and greater than 1 year before depression assessment, respectively. The mean age of the study participants at the time of depression assessment was 56.6 years. 55.1% of participants were female, 89.2% were White, and 71.7% were married. Sixty-eight percent of cases had attained a college education or higher, and 56.1% had a household income >$70,000 USD. Furthermore, 34.5% and 41.5% of the study participants were overweight or obese, respectively; 65.5% reported regular exercise. 47.9% of participants reported alcohol use, 6.2% were currently smoking, and 8.5% reported marijuana use during the 30 days before the depression assessment (Table 1).
Soft tissue sarcomas accounted for 83.2% of cases, while 16.8% of cases comprised sarcomas of the bone. Over half (52.0%) occurred in an extremity. In the past 30 days before the depression assessment, 15.7% of patients had received any modality of treatment, 5.1% of participants had received antidepressants, and 5.2% had received analgesic medications. Patients who received their most recent cancer treatment during the 1-year period before completing the depression assessment were more likely to receive analgesics compared with those who did not receive treatment or those who received their last treatment greater than 1 year before the assessment. Additionally, participants who underwent cancer treatment did not differ from those who did not with respect to demographic characteristics, clinical features, or lifestyle factors, and reported emotional support, except for alcohol use and tumor sites (Table 1). Moreover, approximately 31.6% of participants (n = 283) did not complete the PROMIS-57 depression measurement. Incomplete responses occurred more in females and patients between age 40 and 59 years. Participants who did not complete PROMIS-57 were also more likely to decline to answer questions about race, marriage status, educational attainment, and household income (Appendix Table A1, online only).
The mean depression T-score was 48.3 (SD, 10.0); 58.8% had a T-score <50, 27.9% scored between 50 and 59, and 13.2% scored ≥60. No differences were observed by interval of the last previous cancer treatment to depression assessment regarding the eight individual items of the PROMIS-57 depression assessment or the total depression T-scores (Table 2).
Table 3 shows the prevalence of depression T-score levels by selected patients’ demographic characteristics, clinical features, and modifiable factors. A T-score ≥60 was significantly more frequent among younger patients (P = 4.79 × 10^−5^), those with a low household income (P = 1.99 × 10^−4^), no regular exercise (P = 1.20 × 10^−4^), smoking (ever or current), marijuana use, or lower emotional support scores.
Multivariable analyses showed that compared with participants age 60 years and older, participants in the age ranges of 18–39 years and 40–59 years were more likely to have a T-score ≥60. Adjusted ORs (aORs) and 95% CIs for depression T-scores ≥60 were 3.65 (1.70 to 7.83) for those age 18–39 years and 2.80 (1.52 to 5.17) for those age 40–59 years, compared with participants age 60 years and older. Compared with participants with household incomes <70,000–120,000 USD were less likely to have a T-score of 50–59 or ≥60. aORs and 95% CIs for T-scores of 50–59 and ≥60 were 0.52 (0.30 to 0.91) and 0.46 (0.23 to 0.92), respectively, for income of 120,000 USD; and they were 0.57 (0.32 to 1.01) and 0.15 (0.06 to 0.37), respectively, for income >45,000 USD (income of 120,000 USD [β = −.39; 95% CI, −0.61 to −0.16; P = 8.43 × 10^−4^] and income >$120,000 USD [β = −.56; 95% CI, −0.80 to −0.31; P = 8.45 × 10^−6^]; Table 4).
Participants who reported marijuana use within the past 30 days were more likely to have a T-score ≥60 (OR, 3.48 [95% CI, 1.46 to 8.27]) and had significantly higher depression scores (β = .36; 95% CI, 0.09 to 0.62; P = .008) compared with nonmarijuana users. Conversely, participants with regular exercise were less likely to have a T-score ≥60 (OR, 0.43 [95% CI, 0.24 to 0.75]) and had significantly reduced depression scores (β = −.16; 95% CI, −0.32 to −0.01, P = .038) compared with those without regular exercise. Additionally, per-SD increase in emotional support scores during the last 7 days was significantly associated with a reduced experience of depression for both a T-score of 50–59 (OR, 0.52 [95% CI, 0.42 to 0.65]) and a T-score ≥60 (OR, 0.37 [95% CI, 0.28 to 0.48]). Per-SD increase in emotional support scores was significantly and linearly associated with decreased depression scores (β = −.37; 95% CI, −0.44 to −0.30; P = 4.68 × 10^–22^; Table 5).
In our cohort of 612 patients with sarcoma, we found that 13.2% demonstrated increased depression symptoms compared with the general population, as evidenced by a T-score ≥60 based on the eight-item PROMIS-57 depression scale. Several demographic characteristics and modifiable factors were associated with depression symptoms, including age, household income, regular exercise, marijuana use, and emotional support.
In 2018, the CDC published data showing that between 2013 and 2016, 7.7%, 8.4%, and 8.0% of individuals in the age ranges 20–39, 40–59, and ≥60 years, respectively, experienced depression.^56^ Our study demonstrated a different trend, with a significant increase in the frequency of T-scores ≥60 among younger patients (19.5% and 15.1% of participants age 18–39 and 40–59 years, respectively), compared with participants older than 60 years (9.2%). We hypothesize that the physical and psychosocial impact of treatment may affect younger more than older patients. Consider that many younger individuals are pursuing higher education and a career, possibly while supporting a family and sensing financial instability. These responsibilities are substantial. Contrast this with older individuals who may have developed positive psychosocial adaptations, demonstrating post-traumatic growth, leading to feelings of hope and optimism.^54,57^ Few studies conducted in the United States have evaluated depression symptoms among current or former patients of all forms of cancer. One such study did find results similar to our own, illustrating significantly lower odds of depression among older patients.^32^ Parallel results are noted among studies evaluating specific cancer types, although many of them are international cohorts.^58–63^ One Japanese cohort study revealed a decrease in depressive symptoms with age, and the frequency of symptoms was greatest among patients age 30–39 years.^33^ Analogous findings were also noted in a study among long-term German survivors of cancer, describing significantly lower depression scores among elderly patients.^40^
To our knowledge, no previous study has evaluated the relationship between household income and symptoms of depression among patients with sarcoma and survivors of sarcoma. We found that a household income >45,000 USD, consistent with observations that lower income levels are associated with depression in the general population. Statistics from the CDC showed that the prevalence of depression decreased among adults at or above 400% of the federal poverty level between 2013 and 2016.^56^ Among the general population, available literature demonstrates lower income is associated with an increased incidence of mental health disorders and depression. These findings are consistent with data representing patients with cancer as a group. In one study evaluating risk factors of depression in a cohort of cancer survivors, household income <$75,000 USD was significantly associated with a diagnosis of depression.^34^
Although the physical and psychological benefits of exercise are well studied, few studies evaluated these benefits among patients with sarcoma or survivors of sarcoma. Physical activity is often recommended as a preventative and therapeutic treatment for depression.^64,65^ How ever, engaging in physical activity can represent a challenge to patients with sarcoma, as sarcomas involve soft tissue and bone, increasing the risk for physical limitations. Within our cohort, participants who engaged in regular exercise had a significantly lower prevalence of T-score ≥60 than those who did not. This appears consistent with previous studies evaluating the effect of exercise on depression in cancer survivors.^66,67^ Future studies following a cohort of patients with sarcoma throughout treatment are needed to understand the nature and causality of the association.
Previous studies have demonstrated an association between cigarette smoking and depression,^68,69^ although its directionality remains unclear.^68^ We found a higher prevalence of depression among tobacco smokers, although not statistically significant, which is consistent with existing data among patients with nonspecific forms of cancer.^30^ Future studies examining the benefits of smoking cessation could help clinicians provide motivation to at-risk patients.
Our data support previous studies’ findings of a positive relationship between marijuana use and depression.^70,71^ What is unknown is whether recreational versus medicinal use influences this association. With the legalization of cannabis products in many states, this requires further research.
We report an inverse relationship between social support and depression, reinforcing previous literature describing similar benefits.^72–75^ To our knowledge, no previous studies have evaluated this relationship among patients with sarcoma. Treatment for many sarcomas can be physically and emotionally challenging. Support from family, friends, and clinicians may provide motivation, resilience, and other tools to help patients adapt. Although more studies are needed to determine the critical elements of emotional support in this setting, our data can provide additional assistance to clinicians when counseling patients and families on the benefits of supportive practices.
Our analysis of depression scores and time intervals since the most recent treatment revealed no clear trend or significant association, consistent with similar studies evaluating the prevalence of depression at diagnostic and treatment intervals. In a study conducted in Portugal, it was shown that no significant difference in depression prevalence was seen among patients with sarcoma in either the diagnostic or treatment phases of care.^76^
A strength of our study is a large study population. Additionally, our patients represent diverse demographic characteristics, including a wide age range, equal sex distribution, and 10% racial minorities. Also, VUMC uses a high-quality EHR system, providing us with a superior resource for data collection and abstraction.
Furthermore, the PROMIS measure was selected as CAUSAL was designed to investigate a range of self-reported outcomes among patients with sarcoma. The PROMIS-57 is a validated and standardized instrument that includes not only depression but also seven other domains, such as anxiety, physical function, pain interference, fatigue, sleep disturbance, and ability to participate in social roles and activities. Although several validated screening tools for depression are available, the PROMIS measure does not ask about somatic symptoms, which may overlap with symptoms of cancer and treatment, influencing screening results.^77^
We identified several limitations of this study. First, our single-center design decreases generalizability. Also, the depression prevalence in our study population could have been underestimated because of a potential selection bias that patients with depression were more likely to be nonrespondent. The exclusion of 283 participants with incomplete responses from the current analysis also reduced the statistical power of our study, rending imprecise estimates of associations. In our study, we found some demographic characteristics were associated with nonresponse to the survey (Appendix Table A1), although they were not specific to depression assessment. Our current study did not address barriers to depression screening directly. No information was collected on the explicit reasons for incomplete surveys. Furthermore, sarcomas represent diverse histologies; not all sarcoma diagnoses, staging, treatments, and prognoses are alike. These differences could influence a patient’s risk of developing symptoms of depression. This descriptive project is based on data from a cross-sectional survey with the primary goal of presenting the prevalence of depression by patient characteristics and its correlates. Our current analysis was not able to take into consideration the stage, number of treatment courses, definitive versus palliative care, previous history of chemotherapy, and comorbidities in the study. Our sample size, although the largest, to our knowledge, evaluating depression among patients with sarcoma, limits the evaluation of individual diagnoses for their associations with depression. We aim to address this in future analyses. We also recognize that our definition of alcohol consumption is broad and does not provide detailed data regarding the amount of alcohol consumption per day or per week, limiting detailed analysis of this variable.
Finally, we could not account for mental health diagnoses before study entry. This information could influence the magnitude of our associations. Future studies able to account for this variable would provide superior results on the degree to which sarcoma diagnosis and treatment are responsible for initiating symptoms of depression versus exacerbating symptoms in patients with a positive history.
Identification of risk factors for depression among specific groups of patients with cancer is critical in developing future intervention strategies. Studies like ours are the first step in addressing the needs of patients with sarcoma, an under-investigated population. The next phase of action is to study how to effectively communicate the risks of depression and the benefits of treatment to high-risk patients. Historically, patients with cancer have demonstrated a reluctance to uptake psychologic services.^78–80^ This highlights the importance of equipping health care providers with cancer-specific adapted information to successfully relate the role of psychologic treatment to their patients.
In conclusion, we have identified important risk factors associated with depression among patients with sarcoma and survivors of sarcoma. These include younger age, decreased household income, infrequent exercise, marijuana use, and inferior emotional support. This knowledge would help clinicians to identify patients with a high risk of developing depression and intervene early in their treatment course. Our study suggests that more resources and research should be prioritized to investigate barriers to getting proper mental health screening and treatment for patients with sarcoma.