Meeting NewsPerspective

Children with AML from poor neighborhoods face twofold greater risk for death

Lena E. Winestone, MD, MSHP
Lena E. Winestone

ORLANDO — ZIP code-based low socioeconomic status appeared associated with an increased risk for mortality among children with acute myeloid leukemia treated on clinical trials, according to study results presented at ASH Annual Meeting and Exposition.

In fact, children from neighborhoods with a low area-based income had a 2.4 times greater risk for early death than children living in areas with middle or high income.

“Although I frankly did not find the significant survival disparities among patients from low-income neighborhoods surprising based on my clinical experience and firsthand caring for patients, the degree to which ZIP code impacts survival and, in particular, early mortality, is nevertheless remarkable,” Lena E. Winestone, MD, MSHP, assistant professor in the department of pediatrics at University of California, San Francisco and pediatric hematologist and oncologist at UCSF Benioff Children’s Hospital told Healio. “Understanding the impact of neighborhood on survival is the first step, but additional studies to investigate the relationship to known racial and ethnic disparities will allow us to begin to mitigate disparities and ensure equitable outcomes across our patient population.”

Because income, education and health insurance coverage are known factors that affect receipt of oncology care, Winestone and colleagues sought to evaluate how area-based measures of socioeconomic status contributed to outcomes disparities among children with AML treated on two Children’s Oncology Group phase 3 clinical trials. Although participation on a clinical trial is though to level access to care, researchers hypothesized that patients from low income and education ZIP codes would have inferior 5-year OS and EFS.

The role of socioeconomic factors in outcomes for pediatric acute lymphoblastic leukemia is well-documented and has largely been linked to oral chemotherapy adherence,” Winestone told Healio. “However, socioeconomic factors have not been evaluated in pediatric AML at the national level or within the Children’s Oncology Group to date. Because the backbone chemotherapy is administered in the inpatient setting, adherence to chemotherapy can largely be removed as a possible contributor to outcome disparities in pediatric AML.”

Of the 2,387 patients enrolled on the AAML0531 and AAML1031 clinical trials, 1,467 met the exclusion criteria for the current analysis and had adequate covariate data.

Researchers assessed ZIP code level median annual household income to define patients’ neighborhoods as being in poverty (< $24,250; n = 27), low ($24,250-$56,516; n = 954) and middle/high (> $56,516; n = 695) income areas. They also assessed ZIP code level educational attainment and insurance coverage.

Overall, black and Hispanic children were more likely than white children to live in poverty, low-income and low-education areas, and they were more likely to have Medicaid-only insurance.

Researchers found that rates of 5-year OS were significantly lower among children living in poverty (43%) and low-income areas (61%) compared with those living in middle/high-income areas (68%; P = .004). They observed a similar trend for 5-year EFS (34% vs. 46% vs. 54%; P = .005).

Patients living in areas of lowest educational attainment also demonstrated poorer rates of 5-year OS (58% vs. 70%; P = .005) and EFS (44% vs. 54%; P = .03) than those living in the areas with the highest educational attainment.

Having Medicaid-only insurance was associated with lower rates of 5-year OS (59 ± 5% vs. 66 ± 3%; P = .01) but similar rates of EFS (48 ± 5% vs. 50 ± 3%).

Even after adjusting for area-based education, insurance and established risk factors, results of a multivariable model showed children from middle/high-income areas had a 21% reduced risk for mortality compared with children from low-income areas (adjusted HR = 0.79; 95% CI, 0.63-0.99). They also demonstrated an improvement in EFS (adjusted HR = 0.77; 95% CI, 0.65-0.89).

However, the differences in survival according to area-based educational attainment and insurance coverage did not persist on the multivariable analysis, suggesting that area-based income and established risk factors may be driving the association, according to the researchers.

Moreover, area-based low income was associated with an increased risk for early death (OR = 2.43; 95% CI, 1.04-5.69) and higher incidence of treatment-related mortality (11.1 ± 10.5% vs. 3.7 ± 2.5%; P = .03) compared with high/middle-income areas.

“Just as race/ethnicity, gender and age are often included in analyses of results, these data suggest that area-based income should also be incorporated into risk models of pediatric AML,” Winestone told Healio. “Moreover, now that we have appended these variables to the Children’s Oncology Group master dataset, they are available for other investigators to incorporate into their analyses. It is my hope that our findings will also lead to the prospective collection of individual socioeconomic factors — such as income, education and employment — through upcoming clinical trials. Collection of these data would allow us to start identifying specific individual patients who would benefit from targeted poverty-directed interventions.”

Such interventions might include providing resources like transportation, nutrition or housing, or offering tailored support from a health navigator who can help with insurance issues, follow-up, low health literacy and lack of trust, Winestone added. Clinicians may also consider changes in medical management of comorbidities and complications that tend to impact patients from low socioeconomic areas.

Moving forward, researchers aim to further understand why these disparities persisted even among patients treated on clinical trials.

“One possible explanation we are currently investigating is potential differences in supportive care because most of the gains in survival in pediatric AML have been related to supportive care rather than benefits from chemotherapy or other AML-targeted therapies,” Winestone said. “Although these clinical trials included in this study contained recommendations for supportive care management, such as antibiotic prophylaxis and screening for cardiotoxicity, these were not mandated by the trial and thus could be a place where care differed.

“In addition, the disparity in early deaths we observed suggests that patients from low-income areas may die before they have the opportunity to benefit from the care received through the clinical trial,” she added. – by Alexandra Todak

Reference:

Winestone LE, et al. Abstract 703. Presented at: ASH Annual Meeting and Exposition; Dec. 7-10, 2019; Orlando.

Disclosures: Winestone reports no relevant financial disclosures. One study author reports research funding from Merck and Pfizer and a data safety monitoring board chair role for an antifungal study by Astellas.

Lena E. Winestone, MD, MSHP
Lena E. Winestone

ORLANDO — ZIP code-based low socioeconomic status appeared associated with an increased risk for mortality among children with acute myeloid leukemia treated on clinical trials, according to study results presented at ASH Annual Meeting and Exposition.

In fact, children from neighborhoods with a low area-based income had a 2.4 times greater risk for early death than children living in areas with middle or high income.

“Although I frankly did not find the significant survival disparities among patients from low-income neighborhoods surprising based on my clinical experience and firsthand caring for patients, the degree to which ZIP code impacts survival and, in particular, early mortality, is nevertheless remarkable,” Lena E. Winestone, MD, MSHP, assistant professor in the department of pediatrics at University of California, San Francisco and pediatric hematologist and oncologist at UCSF Benioff Children’s Hospital told Healio. “Understanding the impact of neighborhood on survival is the first step, but additional studies to investigate the relationship to known racial and ethnic disparities will allow us to begin to mitigate disparities and ensure equitable outcomes across our patient population.”

Because income, education and health insurance coverage are known factors that affect receipt of oncology care, Winestone and colleagues sought to evaluate how area-based measures of socioeconomic status contributed to outcomes disparities among children with AML treated on two Children’s Oncology Group phase 3 clinical trials. Although participation on a clinical trial is though to level access to care, researchers hypothesized that patients from low income and education ZIP codes would have inferior 5-year OS and EFS.

The role of socioeconomic factors in outcomes for pediatric acute lymphoblastic leukemia is well-documented and has largely been linked to oral chemotherapy adherence,” Winestone told Healio. “However, socioeconomic factors have not been evaluated in pediatric AML at the national level or within the Children’s Oncology Group to date. Because the backbone chemotherapy is administered in the inpatient setting, adherence to chemotherapy can largely be removed as a possible contributor to outcome disparities in pediatric AML.”

Of the 2,387 patients enrolled on the AAML0531 and AAML1031 clinical trials, 1,467 met the exclusion criteria for the current analysis and had adequate covariate data.

Researchers assessed ZIP code level median annual household income to define patients’ neighborhoods as being in poverty (< $24,250; n = 27), low ($24,250-$56,516; n = 954) and middle/high (> $56,516; n = 695) income areas. They also assessed ZIP code level educational attainment and insurance coverage.

PAGE BREAK

Overall, black and Hispanic children were more likely than white children to live in poverty, low-income and low-education areas, and they were more likely to have Medicaid-only insurance.

Researchers found that rates of 5-year OS were significantly lower among children living in poverty (43%) and low-income areas (61%) compared with those living in middle/high-income areas (68%; P = .004). They observed a similar trend for 5-year EFS (34% vs. 46% vs. 54%; P = .005).

Patients living in areas of lowest educational attainment also demonstrated poorer rates of 5-year OS (58% vs. 70%; P = .005) and EFS (44% vs. 54%; P = .03) than those living in the areas with the highest educational attainment.

Having Medicaid-only insurance was associated with lower rates of 5-year OS (59 ± 5% vs. 66 ± 3%; P = .01) but similar rates of EFS (48 ± 5% vs. 50 ± 3%).

Even after adjusting for area-based education, insurance and established risk factors, results of a multivariable model showed children from middle/high-income areas had a 21% reduced risk for mortality compared with children from low-income areas (adjusted HR = 0.79; 95% CI, 0.63-0.99). They also demonstrated an improvement in EFS (adjusted HR = 0.77; 95% CI, 0.65-0.89).

However, the differences in survival according to area-based educational attainment and insurance coverage did not persist on the multivariable analysis, suggesting that area-based income and established risk factors may be driving the association, according to the researchers.

Moreover, area-based low income was associated with an increased risk for early death (OR = 2.43; 95% CI, 1.04-5.69) and higher incidence of treatment-related mortality (11.1 ± 10.5% vs. 3.7 ± 2.5%; P = .03) compared with high/middle-income areas.

“Just as race/ethnicity, gender and age are often included in analyses of results, these data suggest that area-based income should also be incorporated into risk models of pediatric AML,” Winestone told Healio. “Moreover, now that we have appended these variables to the Children’s Oncology Group master dataset, they are available for other investigators to incorporate into their analyses. It is my hope that our findings will also lead to the prospective collection of individual socioeconomic factors — such as income, education and employment — through upcoming clinical trials. Collection of these data would allow us to start identifying specific individual patients who would benefit from targeted poverty-directed interventions.”

Such interventions might include providing resources like transportation, nutrition or housing, or offering tailored support from a health navigator who can help with insurance issues, follow-up, low health literacy and lack of trust, Winestone added. Clinicians may also consider changes in medical management of comorbidities and complications that tend to impact patients from low socioeconomic areas.

PAGE BREAK

Moving forward, researchers aim to further understand why these disparities persisted even among patients treated on clinical trials.

“One possible explanation we are currently investigating is potential differences in supportive care because most of the gains in survival in pediatric AML have been related to supportive care rather than benefits from chemotherapy or other AML-targeted therapies,” Winestone said. “Although these clinical trials included in this study contained recommendations for supportive care management, such as antibiotic prophylaxis and screening for cardiotoxicity, these were not mandated by the trial and thus could be a place where care differed.

“In addition, the disparity in early deaths we observed suggests that patients from low-income areas may die before they have the opportunity to benefit from the care received through the clinical trial,” she added. – by Alexandra Todak

Reference:

Winestone LE, et al. Abstract 703. Presented at: ASH Annual Meeting and Exposition; Dec. 7-10, 2019; Orlando.

Disclosures: Winestone reports no relevant financial disclosures. One study author reports research funding from Merck and Pfizer and a data safety monitoring board chair role for an antifungal study by Astellas.

    Perspective
    Ann F. Mohrbacher

    Ann F. Mohrbacher

    These results are generalizable because the researchers derived data from two Children’s Oncology Group trials with patients from cities across the U.S. The question is, what is causing these differences in outcomes? Are these patients not getting to their appointments? Are their parents less educated or less sophisticated? The researchers make the association strictly by socioeconomic status — without an impact of education, insurance and established risk factors — but they don’t necessarily show causality. Usually, patients who are enrolled on clinical trials have better outcomes because there is support in place to make sure patients make their appointments required by the trial protocol.

    Depending on the trial protocols, it would be interesting to know whether the therapies evaluated were oral agents. If so, and patients were taking their medications at home, were they documenting their compliance with the therapy? Is that less sophisticated because of the medical literacy of the family?

    There could be other contributing factors that correlate with socioeconomic status. For instance, obesity and literacy also need to be cosorted from the poverty association. Language barriers tend to be higher among lower socioeconomic groups. Obesity is another factor associated with poverty that can create some biological differences. In studies in acute lymphoblastic leukemia, researchers have shown associations between obesity and increased risk for toxicity and poorer outcomes. What previously was being attributed to certain ethnic groups correlated more with obesity that was more prevalent in that group.

    Usually, many of the barriers we typically see do not persist on clinical trials. For that reason, these data are somewhat surprising. I believe we need to look at other correlates of poverty that might be affecting these outcomes.

    • Ann F. Mohrbacher, MD
    • Keck School of Medicine of USC

    Disclosures: Mohrbacher reports no relevant financial disclosures.

    See more from ASH Annual Meeting and Exposition