Race, Ethnicity, and Socioeconomic Factors as Determinants of Cachexia Incidence and Outcomes in a Retrospective Cohort of Patients With Gastrointestinal Tract Cancer

PURPOSE

Cachexia is a paraneoplastic syndrome of unintentional adipose and muscle tissue wasting with severe impacts to functionality and quality of life. Although health inequities across minority and socioeconomically disadvantaged groups are known, the role of these factors in cachexia progression is poorly characterized. This study aims to evaluate the relationship between these determinants and cachexia incidence and survival in patients with gastrointestinal tract cancer.

METHODS

Through retrospective chart review from a prospective tumor registry, we established a cohort of 882 patients with gastroesophageal or colorectal cancer diagnosed between 2006 and 2013. Patient race, ethnicity, private insurance coverage, and baseline characteristics were evaluated through multivariate, Kaplan-Meier, and Cox regression analyses to determine associations with cachexia incidence and survival outcomes.

RESULTS

When controlling for potentially confounding covariates (age, sex, alcohol and tobacco history, comorbidity score, tumor site, histology, and stage), Black (odds ratio [OR], 2.447; P < .0001) and Hispanic (OR, 3.039; P < .0001) patients are at an approximately 150% and 200%, respectively, greater risk of presenting with cachexia than non-Hispanic White patients. Absence of private insurance coverage was associated with elevated cachexia risk (OR, 1.439; P = .0427) compared to privately insured patients. Cox regression analyses with previously described covariates and treatment factors found Black race (hazard ratio [HR], 1.304; P = .0354) to predict survival detriments, while cachexia status did not reach significance (P = .6996).

CONCLUSION

Our findings suggest that race, ethnicity, and insurance play significant roles in cachexia progression and related outcomes that are not accounted for by conventional predictors of health. Disproportionate financial burdens, chronic stress, and limitations of transportation and health literacy represent targetable factors for mitigating these health inequities.

INTRODUCTION

Race, ethnicity, and socioeconomic factors are heavily correlated with poor health outcomes in the oncologic field,¹ with accumulating evidence indicating this impact extends to cancer cachexia. Cancer cachexia is a paraneoplastic state defined as weight loss of 5% or greater within the 6 months leading up to a cancer diagnosis² and predisposes patients to greater morbidity and mortality. Cachexia is understood to result from the interplay of metabolic and inflammatory upregulation driven by tumor-induced systemic changes in basal energy utilization and expenditure.^3,4 Cachexia is classically recalcitrant to conventional therapeutic options within any given cancer subtype. Despite its profound consequences on survival and quality of life, there is limited understanding of cachexia incidence and prevalence within the racial and socioeconomic framework.

CONTEXT

• Key Objective

• To determine whether patient demographic and socioeconomic factors predict for cachexia incidence and cachexia-associated survival outcomes in a large retrospective cohort of patients with GI tract cancer.

• Knowledge Generated

• Patients who were Black, Hispanic, or lacked private insurance coverage were significantly more likely to meet criteria for cachexia at the time of their initial cancer diagnosis. Survival detriments associated with cancer cachexia were consistently observed across racial ethnic groups.

• Relevance

• As strategies for detecting and treating cancer cachexia develop, the mechanisms contributing to outcome disparities must simultaneously be addressed. In addition to emphasizing the importance of equitable healthcare access and care continuity, our findings highlight the potential benefit of clinical care protocols and health literacy initiatives designed to promote consistent detection of unintentional weight loss in disadvantaged populations.

Although socioeconomic status is often understood as a gradient, it is crucial to unravel its various dimensions to identify which factors contribute to the development of disparate health outcomes. Understanding the specific causative factors behind disproportionate health burden is essential for targeted revisions in health care policy.

Through the development of a retrospective clinical data set of 3,180 patients with gastrointestinal and pulmonary cancer, we have previously evaluated associations between the incidence of cachexia with pretreatment patient medication use, laboratory values, and positron emission tomography imaging in a search for potential diagnostic and interventional tools against this syndrome.^5-7 In these projects, we have consistently detected a significant influence of patient demographic characteristics on cachectic weight loss at the time of cancer diagnosis. Consequently, this study intends to further characterize the relationship between cachexia and patient race, ethnicity, and socioeconomic factors. Ultimately, we hope to reveal targets for actionable change within physician-patient interactions and the broader health care frameworks they take place in.

METHODS

All statistical analyses were conducted at the 5% significance level using IBM SPSS Statistics for Macintosh, Version 28.0 (Armonk, NY).

Patient Cohort

Through support from the UT Southwestern clinical data registry and chart review for supplemental collection and validation, we extracted patient data when sufficient patient and tumor characteristics were available for meaningful interpretation. Through this collection, we established a database of 952 patients with primary gastroesophageal or colorectal cancer diagnosed between January 1, 2006, and December 31, 2013. Comorbidities were interpreted through the Charlson comorbidity index.⁸ Private insurance was defined as patients having any component of private insurance in their method of payment coverage for their treatment. Patients who were self-pay or relied solely on government-provided insurance, such as Medicare, Medicaid, and Tricare (for military members and veterans), were classified as not having private insurance. Treatment strategy was classified as treatment with palliative intent, treatment with curative intent, or as not receiving cancer-directed therapies. Patients who were Asian (n = 50) or of unknown (n = 20) race were excluded from the analysis because of insufficient representation within the cohort, resulting in a final cohort size of 882 patients. Extraction of data pertaining to cohort characteristics was conducted by authors blinded to patients' cachexia status at diagnosis.

Determination of Cachexia Status at Diagnosis

Cachexia status at diagnosis was retrospectively determined by applying the international consensus definition established by Fearon et al.⁹ In the 6 months before diagnosis, the threshold for unintentional weight loss was 5% to meet cachexia criteria for patients with a BMI ≥ 20. For patients with a BMI < 20, this threshold was 2%. The clinical documentation collected to determine this included quantitative weight logging as well as physician and dietitian documentation.

Multivariate Analysis

Multivariate logistic regression was conducted including patient and tumor characteristics to determine which factors independently contributed to cachexia at diagnosis. Patient covariates included age at diagnosis, sex, race, private insurance status, alcohol use history, tobacco use history, and Charlson comorbidity index score. Tumor covariates included tumor site (gastroesophageal or colorectal), histology, and stage.

Survival Analysis

The Kaplan-Meier estimator was used to calculate and depict survival differences between race/ethnic groups stratified by cachexia status at diagnosis. Survival was further evaluated through Cox regression with cachexia included as a covariate in addition to the covariates used in the multivariate logistic regression. The end point of this analysis was the survival time in months to the last date of contact (censored) or the death date. Hazard ratio (HR) survival plots were generated from this analysis for race/ethnicity category to visualize results.

Comparison of Location-Associated Characteristics

To extract location-associated race, ethnicity, and socioeconomic characteristics, the 2013 5-year American Community Survey (ACS) by the US Census Bureau was used.¹⁰ Zip codes from patient addresses were matched with the following estimates: high school graduation rate for populations age 45-64 years and older than 65 years, the percentage of the population below the poverty level, the percentage of households without vehicles, as well as the percentages of populations that were White (non-Hispanic), Black, or Hispanic of any race. T-tests were conducted to determine significant differences between these factors on the basis of cachexia status at cancer diagnosis. To determine which factors could be evaluated under the assumption of equal variances, Levene's test for equality of variance was utilized with a significance level of P < .05.

RESULTS

Cohort Characteristics and Cachexia Incidence

Table 1 demonstrates the population characteristics of our cohort. Incidence of cachexia across the total cohort was 37.19%. Relative to non-Hispanic White patients, Black and Hispanic patients demonstrated respective increases of 66.98% and 72.75%. Patients without private insurance were found to have a 37.42% relative increase in cachexia incidence compared with patients with any private insurance as part of their medical coverage.

TABLE 1.

Population Characteristics

Race and Private Insurance Predict Cachexia Status at Diagnosis

Table 2 demonstrates the multivariate logistic regression analysis using the covariates: age, sex, race, private insurance, alcohol and tobacco history, Charlson comorbidity index, tumor site, histology, and stage. Black race was found to have an odds ratio (OR) of 2.477 (95% CI, 1.620 to 3.697; P < .0001) toward cachexia incidence at diagnosis. Similarly, Hispanic ethnicity demonstrated an OR of 3.039 (95% CI, 1.943 to 4.754; P < .0001). Patients lacking private insurance had an OR of 1.439 (95% CI, 1.012 to 2.045; P = .0427) toward cachexia incidence.

TABLE 2.

Multivariate Analysis of Cachexia Incidence at Cancer Diagnosis

Black Race Predicts Greater Survival Detriment Than Cachexia Status

Figure 1 demonstrates Kaplan-Meier survival plots by cachexia status at diagnosis, race/ethnicity, and groups of race and ethnicity stratified by cachexia status at cancer diagnosis with associated median survival times. Log-rank survival comparisons found statistically significant differences between any groups with differing cachexia status, without significant differences between groups of differing race or ethnicity with equivalent cachexia status. From this interpretation of survival, cachexia status at diagnosis appeared to be a primary determinant of survival detriment (Fig 1).

FIG 1.

Kaplan-Meier survival by (A) cachexia status at diagnosis, (B) race/ethnicity, and (C) combined stratification. OS, overall survival.

Cox regression analysis included the covariates age, sex, race, private insurance, alcohol and tobacco history, Charlson comorbidity index, tumor site, histology, stage, treatment intent, and cachexia status at cancer diagnosis. This analysis observed significant survival detriments associated with age at diagnosis (HR, 1.024; 95% CI, 1.015 to 1.033; P < .0001), Black race (HR, 1.304; 95% CI, 1.018 to 1.670; P = .0354), gastroesophageal tumor site (HR, 2.492; 95% CI, 1.982 to 3.132; P < .0001), unknown or mixed histology (HR, 1.488; 95% CI, 1.066 to 2.077; P = .0196), and stage (HR, 1.683; 95% CI, 1.491 to 1.900; P < .0001). Curative (HR, 0.258; 95% CI, 0.177 to 0.376; P < .0001) and palliative treatment intent (HR, 0.528; 95% CI, 0.342 to 0.814; P = .0039) were associated with protective effects on overall survival. Hispanic ethnicity (P = .7812), lacking private insurance (P = .7483), and cachexia status at diagnosis (P = .6996) did not reach significant contributions to survival outcomes.

Socioeconomic Factors for Patient Zip Codes Associate With Cachexia

Appendix Table A1 (online only) demonstrates T tests comparing patients with and without cachexia at diagnosis by zip code characteristics, including high school graduation rates, racial and ethnic population composition, poverty level, and lack of vehicle access. For every factor included, increased socioeconomic burden and minority populations were associated with the presence of cachexia at diagnosis. Demographic data corresponding to cachectic patient zip codes demonstrated increased proportions of Black and Hispanic populations compared with demographic data from noncachectic patient zip codes. On average, patients found to have cachexia at diagnosis were significantly more likely to reside in zip codes with lower high school graduation rates, greater poverty incidence, and decreased vehicle access.

DISCUSSION

Understanding the social determinants influencing cancer cachexia is important in therapeutically targeting improved outcomes among disadvantaged groups. In this study, we interrogated tumor registry and patient chart data with the goal of revealing targets for intervention. Known risk factors for cachexia risk include advanced malignancy, aggressive tumor types, advanced age,¹¹ Charlson comorbidity index,¹² and alcohol/tobacco use.^13,14 When we controlled for each of these factors, Black and Hispanic patients still demonstrated more than double the risk of presenting with cachectic weight loss at the time of cancer diagnosis compared with non-Hispanic White patients. A lack of private insurance also held significance in the development of cancer cachexia with a risk elevation approaching 50%. Fundamentally, we observed sustained elevation of cachexia incidence associated with Black race, Hispanic ethnicity, and insurance classification independent of other known cachexia risk factors. Although the causal link for this difference cannot be confirmed by our analysis, we suggest several potential mechanisms for these findings.

Advanced disease is an important risk factor for development of cancer cachexia. Differences in cachexia incidence might be in part attributable to delayed detection of cancer-presenting symptoms. Prior studies have shown that Black and Hispanic patients are less likely to have routine follow-up with primary care providers than non-Hispanic White patients,^15,16 resulting in differences in utilization of screening that places minorities at higher risk for delayed cancer detection.^17-19 However, it is important to note that disparities in cachexia incidence among Hispanic and Black patients persisted despite controlling for tumor stage, indicating that delayed detection is not the sole contributor for these differences.

Several race-associated hypotheses have highlighted the effects of chronic stressors on individuals in minority communities, namely the weathering hypothesis.²⁰ This theory posits that the stressors disproportionately faced by Black and Hispanic populations (poverty, discrimination, and housing insecurity)^21,22 result in elevated levels of the stress hormone cortisol and continued stimulation of sympathetic activity. Eventually, these systems become strained and inefficient, and the loss of cortisol's anti-inflammatory activity results in increased oxidative stress and inflammation.²³ Notably, dysregulation of the sympathetic nervous system is becoming increasingly implicated in cachexia pathogenesis.^2,24 Measures reducing these stressors may have a role in reducing disparity in cachexia development.

Importantly, patient zip code was found to be significantly associated with cachexia incidence. Patients with increased cachexia burden were more likely to reside in zip codes with more residents living below the poverty level. This parallels our multivariate analysis findings indicating greater cachexia risk for patients lacking private insurance. These observations underscore deficits in the public insurance coverage that defines the bottom line of health care access for financially disadvantaged families.

Our zip code analyses found evidence that elevated cachexia incidence is associated with regions where a lower percentage of households had access to a vehicle. Hispanic and Black patients are consistently found at a higher risk of missing medical appointments,²⁵ with one of the most often cited reasons being transportation barriers.²⁶ In Dallas, Texas, Black and Hispanic households are much less likely to have access to a vehicle than non-Hispanic White households.²⁷ These findings suggest the importance of supporting improved and affordable public transportation options for health care accessibility.

In addition to equitable primary care access among racial and ethnic minorities, it is important that low rates of health literacy be addressed. Health literacy is defined as the extent to which patients comprehend basic health processes and health care navigation, which is essential for timely detection of concerning symptoms and seeking of appropriate health care. In the context of cancer and cachexia, adequate health care literacy is crucial for following the guidance of research-based health recommendations^28,29 and recognition of clinically significant symptoms such as unexplained weight loss. It has previously been noted that members of minority communities have lower health literacy,^30,31 a disparity that is in part attributed to factors such as limited educational opportunities and mistrust of health care institutions.^32,33 Assessment of our data in conjunction with the US Census Bureau's ACS displays significantly decreased high school graduation rates, a potential indicator of health literacy in patient zip codes. Addressing disparities in education is key not only in increasing patient ability in gaining access to health care but also allows an improved ability to recognize and monitor for clinical deteriorations at home. Targeted interventions to improve health literacy have already been shown to reduce health disparity^34-36 and should be a primary target in improving early identification of cancer-presenting symptoms such as unintentional weight loss.

Preliminary survival estimation by Kaplan-Meier analysis suggested that cachexia at the time of diagnosis is associated with greater mortality risk, consistent with prior studies correlating cachexia with poor prognosis.^3-4 Of note, there was no clear difference in mortality appreciated between race/ethnicity once cachexia had developed. However, the increased proportion of Black and Hispanic patients who presented with cachexia at diagnosis suggests that these groups are at higher risk for cachexia-associated mortality. Importantly, on multivariate analysis per COX regression, risk of death was found to be 30% higher among Black patients, even when controlling for variables including age, lack of private insurance, tobacco/alcohol history, Charlson comorbidity index, tumor characteristics, and cachexia status at diagnosis.

The observations of this study should be considered in the context of several data set limitations. Cachexia was defined solely by its principal manifestation of unintentional weight loss because of inconsistent testing and documentation of pertinent functional and biochemical parameters. Furthermore, our survival analyses are especially vulnerable to variable interactions and multicollinearity from patient cachexia status, race/ethnicity, treatment course, and tumor characteristics.

Unintentional weight loss is highly specific for cancer,³⁷ and warrants rapid investigation by a health care provider if detected. Black, Hispanic, and socioeconomically disadvantaged populations are at greater risk of cachexia and related survival detriments that are poorly accounted for by typical health determinants. The identification and evaluation of unintentional weight loss can be improved in these groups by addressing limitations in care continuity, health literacy, and public insurance coverage. It is important to improve health literacy with targeted public health messaging educating at-risk communities on common gastrointestinal cancer symptoms, especially in recognizing unintentional weight loss. Improving accessibility of care for vulnerable populations would support earlier detection and intervention invaluable toward preventing functional deterioration from cachexia. This requires identifying and systemically overcoming financial and transportation barriers specific to communities. As research advances cachexia diagnostic and management strategies, we must simultaneously investigate and mitigate unjust socioeconomic factors contributing to preventable burden.

APPENDIX

TABLE A1.

T-Tests Comparing Characteristics Associated With Patient Zip Codes Between Patients With Gastroesophageal Tract Cancer With and Without Cachexia at the Time of Cancer Diagnosis

AUTHOR CONTRIBUTIONS

Conception and design: Santiago Olaechea, Puneeth Iyengar, Rodney Infante

Financial support: Puneeth Iyengar, Rodney Infante

Administrative support: Puneeth Iyengar

Provision of study materials or patients: Puneeth Iyengar, Rodney Infante

Collection and assembly of data: Santiago Olaechea, Linda Anne Gilmore, Christian Alvarez, Puneeth Iyengar

Data analysis and interpretation: Santiago Olaechea, Brandon Sarver, Alison Liu, Puneeth Iyengar, Rodney Infante

Manuscript writing: All authors

Final approval of manuscript: All authors

Accountable for all aspects of the work: All authors

AUTHORS' DISCLOSURES OF POTENTIAL CONFLICTS OF INTEREST

Race, Ethnicity, and Socioeconomic Factors as Determinants of Cachexia Incidence and Outcomes in a Retrospective Cohort of Patients With Gastrointestinal Tract Cancer

The following represents disclosure information provided by authors of this manuscript. All relationships are considered compensated unless otherwise noted. Relationships are self-held unless noted. I = Immediate Family Member, Inst = My Institution. Relationships may not relate to the subject matter of this manuscript. For more information about ASCO's conflict of interest policy, please refer to www.asco.org/rwc or ascopubs.org/op/authors/author-center.

Open Payments is a public database containing information reported by companies about payments made to US-licensed physicians (Open Payments).