Abstract
Purpose:
Area deprivation index (ADI), a measure of neighborhood socioeconomic disadvantage, has been linked to metabolic outcomes in the general population, but has received limited attention in survivors of childhood acute lymphoblastic leukemia (ALL), a population with high rates of overweight and obesity.
Methods:
We retrospectively reviewed heights and weights of ≥5 year survivors of pediatric ALL (diagnosed 1990–2013). Residential addresses were geocoded using ArcGIS to assign quartiles of ADI, a composite of 17 measures of poverty, housing, employment, and education, with higher quartiles reflecting greater deprivation. Odds ratios (OR) and 95% confidence intervals (CI) for the association between ADI quartiles and overweight/obesity or obesity alone were calculated with logistic regression.
Results:
On average, participants (n=454, 50.4% male, 45.2% Hispanic) were age 5.5 years at diagnosis and 17.4 years at follow-up. At follow-up, 26.4% were overweight and 24.4% obese. Compared to the lowest ADI quartile, survivors in the highest quartile were more likely to be overweight/obese at follow up (OR = 2.33, 95% CI: 1.23–4.44) after adjusting for race/ethnicity, sex, age at diagnosis, and age at follow-up. The highest ADI quartile remained significantly associated with obesity (OR = 5.28, 95% CI: 1.79–15.54) after accounting for weight status at diagnosis.
Conclusions:
This study provides novel insights into possible social determinants of health inequalities among survivors of childhood ALL by reporting a significant association between neighborhood deprivation and overweight/obesity.
Implications for Cancer Survivors:
Survivors of childhood ALL residing in neighborhood with greater socioeconomic disadvantage may be at increased risk of overweight and obesity and candidates for targeted interventions.
Keywords: Childhood cancer, cancer survivorship, social determinants of health, neighborhood deprivation, obesity, cancer late effects
INTRODUCTION
With modern therapeutic regimens, long-term survival for pediatric acute lymphoblastic leukemia (ALL), the most common malignancy in childhood, has reached approximately 90% [1]. As the number of pediatric ALL survivors continues to grow, characterizing the long-term health of this population has become a research priority. Chronic health conditions are highly prevalent among adult survivors of childhood and adolescent cancers [2, 3], with long-term survivors of pediatric ALL particularly vulnerable to adult overweight and obesity. [4]. Given that obesity contributes to adverse cardiometabolic profiles and cardiovascular risk in the general population [5, 6], and survivors of childhood ALL experience elevated rates of adverse cardiovascular outcomes [7, 8], it is imperative to develop and apply obesity interventions in childhood ALL survivors. Effective intervention strategies for ALL survivors requires a clear understanding of those who are at greatest risk of becoming obese.
Exposure to CRT is one of the few well-established risk factors for overweight and obesity among survivors of childhood ALL [9]. However, the frequency of CRT has decreased dramatically in recent decades in favor of central nervous system-directed chemotherapy. Notably, treatment-related obesity has persisted across decades of evolution in treatment strategies and is prevalent regardless of age at diagnosis, sex, or treatment exposure [4, 9–11]. Pediatric patients with ALL exhibit deficits in physical fitness and endurance shortly after the initiation of multi-agent chemotherapy [12]. Correspondingly, a number of studies have demonstrated that ALL patients treated with chemotherapy alone experience rapid weight gain while on therapy [11], with a proportion of patients retaining excess adiposity well beyond the end of treatment [10]. Because clinical and demographic factors only account for a fraction of the observed variability in post-treatment adiposity, consideration should be given to alternative factors potentially involved in the development of obesity in survivors of pediatric ALL treated with contemporary chemotherapy. In the general population, studies have linked various indicators of socioeconomic status (SES) to obesity prevalence [13–15]. While most studies examining the association between SES and obesity have focused on individual-level factors [16–18], there is now increasing epidemiologic focus on developing a systems-oriented multilevel framework for addressing the global obesity epidemic. In particular, area-based socioenvironmental factors, which may restrict an individual’s access to healthcare and participation in a healthy lifestyle [19], have been linked to obesity in the general population[20, 21], survivors of adult malignancies[22], and, most recently, adult survivors of childhood cancers[23].
Singh et al. developed and validated a composite summary of 17 measures of neighborhood socioeconomic disadvantage for the US—the area deprivation index (ADI) [24]. This ADI captures information on area-based poverty, housing, employment, and education to provide insight into social determinants of health inequalities. ADI has been associated with disease risk factors [25], healthcare utilization [26], and disparities in health outcomes [27]. Specifically, area deprivation appears to directly correlate with the prevalence of overweight and obesity in the general population [28]. Nonetheless, possible links between ADI and obesity specifically among survivors of pediatric ALL are yet to be examined. Therefore, the objective of the current study was to evaluate the association between ADI and overweight and obesity among a diverse population of survivors of childhood ALL, which could point to novel risk factors for this important chronic health condition among these individuals.
METHODS
Study setting, participant recruitment and enrollment.
Eligible participants included individuals who: 1) were diagnosed with ALL at Texas Children’s Hospital (TCH) in Houston, Texas between 1990 and 2013; 2) survived at least five years from diagnosis; 3) had heights and weights at long-term survivor visit recorded in their health record; and 4) had a residential address in the state of Texas at their last clinical appointment. Individuals treated with CRT and those who received bone marrow transplant were excluded due to the potential impact of these therapeutic exposures on growth hormone levels and cardiometabolic outcomes. Similarly, individuals with pre-existing conditions (e.g., Down syndrome, dwarfism) that potentially impacted growth trajectories were excluded. Our final analysis included 454 eligible participants. The study procedures were reviewed and approved by the Baylor College of Medicine Institutional Review Board.
Data collection.
Our data came from two main sources. First, for the individual-level data (including residential information), we relied on patient data that were identifiable in the electronic health record (EHR) system at TCH. Second, data representing neighborhood-level measures were retrieved from the United States Census Bureau (U.S. Census). Specifically, we used the U.S. Census American Community Survey (ACS) 5-year estimates data that were summarized to the census tract geographic level (i.e., neighborhood level). The ACS is a nationwide survey of over 3.5 million households across the U.S. that collects and produces information on social, economic, housing, and demographic characteristics about U.S. population every year [29]. To properly align individual-level data with neighborhood-level data, we used two non-overlapping ACS versions. The 2007–2011 ACS estimates were matched with the individual-level data collected between 2003 and 2011 (n=90) while the 2012–2016 ACS estimates were matched with individual-level data collected between 2012 and 2016 (n=364) (Figure 1).
Figure 1.
Distribution of last available BMI for research participants and corresponding census data from American Community Survey (ACS). Individual-level data for COHORT members from 2003 to 2006 (n=21) were matched with 2007–2011 ACS data set. This was considered acceptable for the current analysis because: (1) neighborhood-level characteristics (from ACS) may not have changed drastically during the time periods under consideration and (2) The first ACS 5-year Estimates version by the U.S. Census Bureau was the 2005–2009 version, and using this version at all would have allowed our ACS data sets to overlap in time – a situation we wanted to avoid.
Outcome: weight status.
Participants’ weight status was specified in terms of body mass index (BMI) in kg/m2, based on clinically recorded information in the TCH EHR. For participants less than 20 years old, we categorized weight according to the 2000 CDC age and weight specific BMI percentiles as follows: underweight <5%, normal weight 5–84.9%, overweight 85–94.9%, obese ≥95% [30]. For those above 20 years, we used the following categories based on the calculated BMI (kg/m2): <18.5 (underweight), 18.5–24.9 (healthy weight), 25–29.9 (overweight), ≥30 (obese). Individuals who were underweight at follow up were excluded from statistical comparisons. Weight status was examined in two ways; as “overweight or obese vs. normal weight” and “obese vs. normal weight.”
Primary Exposure: Area Deprivation Index (ADI).
The ADI is a composite measure of “neighborhood” socioeconomic disadvantage that is based on 17 U.S. Census measures from the following four categories: poverty, housing, employment, and education. Data to compute ADI were retrieved from the ACS 5-year estimates [29]. We followed Singh’s formula, described elsewhere [24], to compute ADI scores for all the census tracts in the state of Texas. For analysis purposes, we classified the ADI data into quartiles (Q1 though Q4) based on the distribution of scores in the state of Texas. Each participant’s last available residential address was geocoded using the ArcGIS World Geocoding Service in ArcGIS Pro 2.2 (Esri, Redlands, CA). Finally, we assigned the ADI score and the quartile-based class of any census tract to the participants whose residential addresses were located inside the tract, with higher quartiles reflecting higher levels of area-based deprivation. Detailed data on the ADI indicators that were used for the current analysis are presented in the supplementary files. See Table S1 and Figure S1.
Covariates.
Additional covariates were abstracted from participants’ EHRs, including year of diagnosis, age at diagnosis, and age at last available BMI. In all cases, time (years) was treated as a continuous variable for analysis purposes. Sex was dichotomized into female and male. Participants self-reported their race/ethnicity, and we categorized them into four groups: non-Hispanic black, non-Hispanic white, Hispanic, and other (which included those that did not specify any race/ethnicity).
Statistical Analysis
All data were reviewed for errors and recoded as appropriate for the current analysis. Our initial data exploration included: determining the distributions of outcome measures and assessing whether data transformations or discretization were needed, assuring that the underlying assumptions of statistical analyses were satisfied. Using chi-square test for categorical variables and t-test for continuous variables, we examined the bivariate relationships between the current weight status (dichotomized) and the primary independent measure (ADI) as well as each of the potential covariates. Afterwards, variables that showed significant bivariate relationship with weight status were retained for further analysis. We used three regression models to assess the associations between the dichotomous measure of weight status and categorical measures of ADI. Model 1 included ADI as independent variable, while Model 2 included ADI and race/ethnicity, and Model 3 additionally included weight status at diagnosis. All models were adjusted for gender, age at diagnosis, and age at last follow-up as covariates. Observations with missing or incomplete data on any covariate included in the multivariable model were excluded from the analysis. The likelihood ratio test was conducted to evaluate whether any observed relationships (odds ratios) were statistically significant. Statistical packages used for data visualization and statistical analysis included SPSS (Version 24.0, SPSS, Inc., Chicago, IL, USA) and Stata (Release 11.0, STATA Corp., College Station, TX, USA). To evaluate potential collinearity in multivariable models, we calculated a variance inflation factor (VIF) for each covariate included in the model but did not detect strong evidence of collinearity (e.g., VIF <2 for all covariates). All hypothesis tests were two-sided with the level of significance set at p≤0.05. We adjusted for the clustering of individuals within census tract by using the Stata cluster command ‘vce’ (cluster clustvar) option to obtain a robust variance estimate that adjusts for within-cluster correlation [31].
RESULTS
Study Population
Overall, data was abstracted on 460 survivors of pediatric ALL, of which six were excluded because they were underweight at last follow up. The clinical and demographic characteristics of the eligible study population (n=454) are presented in Table 1. Participant age at diagnosis ranged from 0.24 years to 16.88 years (mean = 5.53 years, standard deviation [SD] = 3.65); age at last follow-up ranged from 6.95 years to 42.04 years (mean = 17.36 years, SD = 5.16). Participants were mostly male (50.5%), non-Hispanic white (44.7%) or Hispanic (45.3%), and diagnosed with pre-B ALL (94.9%). Overall, 23.8% (n=80) of eligible participants were overweight or obese at the time of their ALL diagnosis.
Table 1:
Key characteristics of study participants by weight status (N=454).
| Normal Weight1 | Overweight or Obese | Obese | |||
|---|---|---|---|---|---|
| Variables | (n=223) | (n=231) | P 2,3 | (n=111) | P 2,4 |
| Gender, n(%) | 0.513 | 0.561 | |||
| Male | 109 (48.9) | 120 (52.0) | 58 (52.3) | ||
| Female | 114 (51.1) | 111 (48.0) | 53 (47.8) | ||
| Race / Ethnicity, n(%) | 0.001 | 0.001 | |||
| N.H. White | 119 (53.4) | 84 (36.4) | 35 (31.5) | ||
| Hispanic | 82 (36.8) | 123 (53.3) | 59 (53.2) | ||
| N.H. Black | 10 (4.5) | 16 (6.9) | 12 (10.8) | ||
| N.H. Other | 12 (5.4) | 8 (3.5) | 5 (4.5) | ||
| Overweight/Obese at diagnosis, n(%)5 | <0.001 | <0.001 | |||
| No | 147 (91.8\3) | 109 (62.3) | 36 (41.4) | ||
| Yes | 14 (8.7) | 66 (37.7) | 51 (58.6) | ||
| Year of diagnosis, n(%) | 0.856 | 0.591 | |||
| 1990–1999 | 67 (30.0) | 75 (32.5) | 29 (26.1) | ||
| 2000–2009 | 136 (61.0) | 136 (58.9) | 74 (66.7) | ||
| 2010–2013 | 20 (9.0) | 20 (8.7) | 8 (7.2) | ||
| Diagnosis, n(%)5 | 0.679 | 0.720 | |||
| Pre-B ALL | 195 (94.7) | 213 (95.1) | 106 (96.4) | ||
| Pre-T ALL | 4 (1.9) | 6 (2.7) | 2 (1.8) | ||
| Infantile ALL | 7 (3.4) | 5 (2.2) | 2 (1.8) | ||
| Mean age at diagnosis, years (SD) | 5.40 (3.58) | 5.66 (3.71) | 0.454 | 5.74 (3.56) | 0.423 |
| Mean age at last follow up, years (SD) | 17.24 (5.39) | 17.56 (4.95) | 0.519 | 17.62 (5.19) | 0.545 |
| ADI Quartiles, n(%) | 0.001 | <0.001 | |||
| Quartile 1 | 88 (39.5) | 51 (22.1) | 18 (16.2) | ||
| Quartile 2 | 51 (22.9) | 63 (27.3) | 27 (24.3) | ||
| Quartile 3 | 54 (24.2) | 67 (29.0) | 35 (31.5) | ||
| Quartile 4 | 30 (13.4) | 50 (21.6) | 31 (27.9) | ||
Underweight individuals (N=6) excluded from analysis
Bivariate relationship examined with chi-square test for categorical variables and t-test for continuous variables. Tests are two-sided with the level of significance set at p≤0.05
Weight status dichotomized as “Normal Weight” vs. “Overweight or Obese”
Weight status dichotomized as “Normal Weight” vs. “Obese”
Data missing for overweight or obese at diagnosis (n=118) and diagnosis (n=24)
Odds of Overweight or Obese Based on Neighborhood (Census Tract) Area Deprivation Index (ADI)
As shown in Table 1, individuals who were overweight or obese were more likely to live in areas characterized by higher deprivation compared to those who were normal weight (e.g., 21.6% vs, 13.4%, p=0.001). When participants who were normal weight were compared with those who were either “overweight or obese” or those “obese only,” only ADI, race/ethnicity (p=0.001) and weight status at diagnosis (p<0.001) differed significantly between the groups (Table 1). Table 2 shows data on the likelihood of being overweight or obese based on the ADI class of participants’ neighborhoods in three multivariable regression models. For Model 1, higher ADI quartile (higher levels of area deprivation) is associated with higher proportion of overweight or obese survivors. This relationship remained statistically significant after accounting for race/ethnicity (Model 2). In Model 2, compared to non-Hispanic whites, the odds of being overweight or obese was significantly higher for Hispanic (OR=1.65, 95% CI: 1.07–2.57) survivors. However, the associations between overweight/obesity and ADI and race/ethnicity were attenuated slightly and were no longer significant after accounting for weight status at diagnosis in the subset of the population with this information available (Model 3). Compared to models evaluating overweight and obesity combined (Table 2), ADI demonstrated stronger associations with obesity alone (Table 3). Increasing quartiles of ADI were significantly associated with the odds of obesity in Model 1 (adjusting for gender, age at diagnosis, and age at last follow-up) and Model 2 (adjusting for covariates in Model 1 and race/ethnicity). Adjusting for weight status at diagnosis did not significantly impact the observed association between ADI and the odds of obesity at last follow-up (Model 3). For example, compared to individuals in the lowest quartile of ADI, the odds of obesity was significantly higher among individuals in the second (OR=2.95, 95% CI: 1.10–7.93), third OR=3.08, 95% CI: 1.11–8.54), and highest (OR=5.28, 95% CI: 1.79–15.54) quartiles. In fact, when all variables were considered collectively, only ADI and weight status at diagnosis were significantly associated with the odds of obesity at last follow-up. The observed associations between quartiles of each individual component of ADI and overweight or obesity are presented in Supplemental Tables S2 and S3, accounting for race/ethnicity, weight status at diagnosis, gender, age at diagnosis, and age at last follow up.
Table 2:
Relationships (Odds Ratio)a,b between Area Deprivation Index (ADI) and weight status (overweight or obese) among survivors of pediatric ALL
| Variables | n | Model 1 (n=440) |
Model 2 (n=440) |
Model 3 (n=336) |
||||
|---|---|---|---|---|---|---|---|---|
| OR | 95% CI | OR | 95% CI | OR | 95% CI | |||
| Area Deprivation | Quartile 1 | 1.00 | Referent | 1.00 | Referent | 1.00 | Referent | |
| Quartile 2 | 2.23 | 1.32 – 3.76 | 1.95 | 1.14 – 3.33 | 1.77 | 0.94 – 3.36 | ||
| Quartile 3 | 2.21 | 1.32 – 3.69 | 1.85 | 1.08 – 3.17 | 1.74 | 0.89 – 3.39 | ||
| Quartile 4 | 3.09 | 1.69 – 5.63 | 2.33 | 1.23 – 4.44 | 2.00 | 0.92 – 4.28 | ||
| Race/Ethnicity | N.H. White | 1.00 | Referent | 1.00 | Referent | |||
| Hispanic | 1.65 | 1.07 – 2.57 | 1.41 | 0.84 – 2.37 | ||||
| N.H. Black | 1.86 | 0.78 – 4.39 | 1.20 | 0.38 – 3.77 | ||||
| N.H. Other | 0.95 | 0.37 – 2.46 | 0.75 | 0.21 – 2.66 | ||||
| Overweight/Obese at diagnosis | No | 1.00 | Referent | |||||
| Yes | 6.11 | 3.21 – 11.64 | ||||||
Covariates were added to the regression models sequentially. Race/ethnicity alone was added to model 2. Both race/ethnicity and weight status at cancer diagnosis were added to model 3. Models adjusted for variables displayed in the table as well as gender, age at diagnosis, and age at last follow up.
Due to the inherent clustering of participants within census tracts (shared ADI score), The vce(cluster clustvar) command in STATA was used to specify that the standard errors allow for intragroup correlation.
Table 3:
Relationships (Odds Ratio)a,b between Area Deprivation Index (ADI) and weight status (obese only) among survivors of pediatric ALL
| Variables | n | Model 1 (n=323) | Model 2 (n=323) | Model 3 (n=248) | ||||
|---|---|---|---|---|---|---|---|---|
| OR | 95% CI | OR | 95% CI | OR | 95% CI | |||
| Area Deprivation | Quartile 1 | 1.00 | Referent | 1.00 | Referent | 1.00 | Referent | |
| Quartile 2 | 2.97 | 1.45 – 6.08 | 2.63 | 1.25 – 5.50 | 2.95 | 1.10 – 7.93 | ||
| Quartile 3 | 3.57 | 1.78 – 7.15 | 3.09 | 1.49 – 6.43 | 3.08 | 1.11 – 8.54 | ||
| Quartile 4 | 6.19 | 2.86 – 13.39 | 4.97 | 2.17 – 11.39 | 5.28 | 1.79 – 15.54 | ||
| Race/Ethnicity | N.H. White | 1.00 | Referent | 1.00 | Referent | |||
| Hispanic | 1.55 | 0.88 – 2.73 | 1.32 | 0.63 – 2.76 | ||||
| N.H. Black | 2.98 | 1.14 – 7.81 | 2.14 | 0.53 – 8.73 | ||||
| N.H. Other | 1.52 | 0.49 – 4.77 | 1.08 | 0.18 – 6.20 | ||||
| Overweight/Obese at diagnosis | No | 1.00 | Referent | |||||
| Yes | 16.22 | 7.59 – 34.64 | ||||||
Covariates were added to the regression models sequentially. Race/ethnicity alone was added to model 2. Both race/ethnicity and weight status at cancer diagnosis were added to model 3. Models adjusted for variables displayed in the table as well as gender, age at diagnosis, and age at last follow up.
Due to the inherent clustering of participants within census tracts (shared ADI score), The vce(cluster clustvar) command in STATA was used to specify that the standard errors allow for intragroup correlation.
DISCUSSION
The current analysis addresses gaps in obesity research among survivors of childhood cancer by examining the possible effects of social determinants of health. Specifically, we compared the prevalence of overweight and obesity in survivors across levels of ADI and found that, compare to survivors in the lowest quartile of ADI, survivors in the highest quartile of ADI were approximately two to three times more likely to be overweight and five to six times more likely to be obese at follow-up. These findings are consistent with recent studies reporting inverse associations between neighborhood socioeconomic status and excess adiposity in survivors of female breast cancer and a heterogenous sample of adult survivors of childhood malignancies [22, 23]. In this assessment, we demonstrated that residential distress plays a role in chronic health conditions among survivors of ALL treated with contemporary chemotherapy.
To date, most studies of obesity among survivors of childhood ALL have focused on individual-level factors. Social determinants of health (SDoH) are best operationalized as intermediate determinants that are measured at the geographic level. In the current study, we report differences in the odds of overweight and obesity across quartiles of ADI, suggesting overweight and obesity disproportionally affects survivors living in areas of greater residential distress. The role of SDoH in healthcare and health outcomes has received increased attention in recent years [32–34]. SDoH play an important role in healthcare access and quality, disease development and subsequent disease-related morbidity and mortality. A potential approach to understanding neighborhood-level stressors is through the assessment of socioeconomic disadvantage. Residential neighborhood-level socioeconomic disadvantage has both direct and indirect effects on health [35–39]. For example, perceptions of the neighborhood environment as unsafe, violent, or highly disordered can increase feelings of distress, both directly and indirectly, through increased feelings of powerlessness and fear [40, 41]. Differential prevalence of overweight and obesity across levels of ADI may be partially explained by variability in poverty, educational attainment, the built environment, substandard housing, and lack of employment opportunities [42, 43].
Although epidemiologic studies investigating the roles of SDoH on the burden of cancer late effects among childhood cancer survivors are scarce, numerous studies in the general population have reported strong, consistent relationships between socioeconomic factors and obesity [13, 14, 44, 45]. For example, McLaren et al. reported, among other findings, a six-fold increase of obesity prevalence among low-SES individuals compared to those in highest SES group [15]; a finding that was subsequently supported by various independent studies [16–18]. However, since the 2000s, scholars have emphasized the importance of a systems-oriented multilevel framework for addressing the global obesity epidemic. Glass and McAtee [46], whose model integrates biological (genes, cells, and organs) and socioenvironmental (economics, culture, social networks, and features of the physical environment) influences on behaviors, described obesity as an epidemic with roots in complex human behavior, with obvious, but as yet unspecified environmental antecedents. Building upon this model, Huang and colleagues[19] argued for framing obesity as a complex system in which behavior is affected by multiple individual-level factors and socioenvironmental factors, and described socioenvironmental factors as those “related to the food, physical, cultural, or economic environment that enable or constrain human behavior, or both.”[19] Over the last decade, scholars have argued for and validated the importance of examining the impact of “place-based” socioenvironmental factors on health outcomes, including obesity [20, 21]. In this regard, the places where people live, work, and play are frequently considered, though the residential neighborhood enjoys the most attention. Higher levels of area-based deprivation have been associated with excessive pediatric weight gain and higher rates of childhood overweight and obesity in various settings in the general population [47, 48].
This is the first study, to our knowledge, to examine the association between area-level deprivation and overweight and obesity among a large, ethnically diverse, contemporary cohort of childhood ALL survivors. Still, the findings presented in this study reflect the experiences at a single childhood cancer treatment center, and may not be broadly generalizable to survivors of childhood ALL treated at other centers. We compared the demographic and clinical characteristics of survivors included in this analysis to the entire population of >5-year survivors of pediatric ALL diagnosed at our institution between 2005 and 2013 (years which were available in the current institutional electronic medical record system).The eligible patients with heights and weights recorded at long-term survivor visits included in this study represent approximately 60% of the total population of survivors. Although we did not observe statistically significant (p-value >0.05) differences between race/ethnicity, sex, age at diagnosis, and diagnosis (B-lineage vs T-lineage) between the study population and the total population of eligible survivors, we are not able to rule out the possibility that patients lost to follow-up between diagnosis and long-term survivor visits biased our results. Furthermore, this cross-sectional chart review evaluated the association between BMI and ADI using the last known address of participants. As such, the current study did not assess influence of potential time-varying exposures, such as changes in individual neighborhood deprivation over time, which may have resulted in some misclassification of exposure to area deprivation. Similarly, data on health habits, including diet, physical activity, and tobacco use, and individual-level socioeconomic factors were unavailable, including household income or parental education, which may modify the association observed between ADI and obesity [49]. Overweight and obesity are complex phenotypes with multifactorial etiologies likely involving genetic, environmental, and life-style factors. Research suggests that pediatric patients with ALL gain weight rapidly during early phases of therapy [11], with weight status at one-year post-diagnosis being predictive of post-treatment BMI [10]. Therefore, future studies are needed to longitudinally track weight trajectories of pediatric patients with ALL and consider the dynamic roles of individual factors including diet and physical activity, area-based SDoH, and treatment factors which interfere with physical fitness and activity during treatment (i.e., asparaginase, corticosteroids) [12].
Conclusion.
Although some demographic and treatment risk factors for adverse cardiometabolic outcomes have been identified in survivors of childhood ALL, this study provides novel insight into possible social determinants of health inequalities. Specifically, this study demonstrates a significant association between neighborhood deprivation and the prevalence of overweight and obesity among long-term survivors of childhood ALL. Contemporary childhood ALL chemotherapy is associated with rapid weight gain during therapy and an increased likelihood of overweight and obesity among survivors [10, 11]. With five-year survival rates approaching 90% for childhood ALL, an understanding of the multifactorial etiology of adverse cardiometabolic profiles is needed to improve the long-term health and quality of life among survivors. The results of this study may aid in the identification of individuals at greater risk for overweight and obesity and point to the need for targeted interventions aimed at survivors residing in areas of greater deprivation. Additional research is warranted to explore the mechanisms by which area deprivation contribute to obesity, particularly among survivors of childhood ALL. An understanding of these pathways may lead to identification of modifiable factors or opportunities for health promotion interventions to reduce adverse weight gain in this susceptible population.
Supplementary Material
Acknowledgments
Grant support
This research was funded by the National Institutes of Health National Cancer Institute (K07 CA218362 to A.L.B.); Leukemia Research Foundation Hollis Brownstein Research Grants Program (to A.L.B); St. Baldrick’s Foundation Consortium Grant (522277 to P.J.L); and the Cancer Prevention & Research Institute of Texas (CPRIT RP160771).
Data Availability:
The datasets generated during and/or analyzed during the current study are available from the corresponding author on reasonable request.
REFERENCES
- 1.Surveillance E, and End Results (SEER) Program, SEER*Stat Database: Incidence - SEER 9 Regs Research Data, Nov 2018 Sub (1975–2016) <Katrina/Rita Population Adjustment> - Linked To County Attributes - Total U.S., 1969–2017 Counties. Released April 2019, based on the November 2018 submission, National Cancer Institute, DCCPS, Surveillance Research Program. [Google Scholar]
- 2.Hudson MM, et al. , Clinical ascertainment of health outcomes among adults treated for childhood cancer. JAMA, 2013. 309(22): p. 2371–81. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Ward E, et al. , Childhood and adolescent cancer statistics, 2014. CA Cancer J Clin, 2014. 64(2): p. 83–103. [DOI] [PubMed] [Google Scholar]
- 4.Meacham LR, et al. , Body mass index in long-term adult survivors of childhood cancer: a report of the Childhood Cancer Survivor Study. Cancer, 2005. 103(8): p. 1730–9. [DOI] [PubMed] [Google Scholar]
- 5.Krauss RM, et al. , Obesity: impact of cardiovascular disease. Circulation, 1998. 98(14): p. 1472–6. [PubMed] [Google Scholar]
- 6.Rosengren A, et al. , Coronary risk factors, diet and vitamins as possible explanatory factors of the Swedish north-south gradient in coronary disease: a comparison between two MONICA centres. J Intern Med, 1999. 246(6): p. 577–86. [DOI] [PubMed] [Google Scholar]
- 7.Giordano P, et al. , Endothelial dysfunction and cardiovascular risk factors in childhood acute lymphoblastic leukemia survivors. Int J Cardiol, 2017. 228: p. 621–627. [DOI] [PubMed] [Google Scholar]
- 8.Oeffinger KC, et al. , Insulin resistance and risk factors for cardiovascular disease in young adult survivors of childhood acute lymphoblastic leukemia. J Clin Oncol, 2009. 27(22): p. 3698–704. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Oeffinger KC, et al. , Obesity in adult survivors of childhood acute lymphoblastic leukemia: a report from the Childhood Cancer Survivor Study. J Clin Oncol, 2003. 21(7): p. 1359–65. [DOI] [PubMed] [Google Scholar]
- 10.Foster KL, et al. , Weight trends in a multiethnic cohort of pediatric acute lymphoblastic leukemia survivors: A longitudinal analysis. PLoS One, 2019. 14(5): p. e0217932. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Zhang FF, et al. , Growth patterns during and after treatment in patients with pediatric ALL: A meta-analysis. Pediatr Blood Cancer, 2015. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Ness KK, et al. , Skeletal, neuromuscular and fitness impairments among children with newly diagnosed acute lymphoblastic leukemia. Leuk Lymphoma, 2015. 56(4): p. 1004–11. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Monteiro CA, et al. , Socioeconomic status and obesity in adult populations of developing countries: a review. Bulletin of the World Health Organization, 2004. 82: p. 940–946. [PMC free article] [PubMed] [Google Scholar]
- 14.Ball K and Crawford D, Socioeconomic status and weight change in adults: a review. Social science & medicine, 2005. 60(9): p. 1987–2010. [DOI] [PubMed] [Google Scholar]
- 15.McLaren L, Socioeconomic status and obesity. Epidemiologic reviews, 2007. 29(1): p. 29–48. [DOI] [PubMed] [Google Scholar]
- 16.Johnson W, et al. , Education modifies genetic and environmental influences on BMI. PloS one, 2011. 6(1): p. e16290. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17.Della Bella S and Lucchini M, Education and BMI: a genetic informed analysis. Quality & Quantity, 2015. 49(6): p. 2577–2593. [Google Scholar]
- 18.Komulainen K, et al. , Education as a moderator of genetic risk for higher body mass index: prospective cohort study from childhood to adulthood. International Journal of Obesity, 2018. 42(4): p. 866. [DOI] [PubMed] [Google Scholar]
- 19.Huang TT, et al. , A systems-oriented multilevel framework for addressing obesity in the 21st century. Preventing chronic disease, 2009. 6(3). [PMC free article] [PubMed] [Google Scholar]
- 20.Janssen I, et al. , Influence of individual-and area-level measures of socioeconomic status on obesity, unhealthy eating, and physical inactivity in Canadian adolescents. The American journal of clinical nutrition, 2006. 83(1): p. 139–145. [DOI] [PubMed] [Google Scholar]
- 21.Cummins S, et al. , Understanding and representing ‘place’in health research: a relational approach. Social science & medicine, 2007. 65(9): p. 1825–1838. [DOI] [PubMed] [Google Scholar]
- 22.Shariff-Marco S, et al. , Impact of Social and Built Environment Factors on Body Size among Breast Cancer Survivors: The Pathways Study. Cancer Epidemiol Biomarkers Prev, 2017. 26(4): p. 505–515. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23.Howell CR, et al. , Neighborhood effect and obesity in adult survivors of pediatric cancer: A report from the St. Jude lifetime cohort study. Int J Cancer, 2019. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24.Singh GK, Area deprivation and widening inequalities in US mortality, 1969–1998. American journal of public health, 2003. 93(7): p. 1137–1143. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25.Wardle J, et al. , Socioeconomic disparities in cancer-risk behaviors in adolescence: baseline results from the Health and Behaviour in Teenagers Study (HABITS). Preventive medicine, 2003. 36(6): p. 721–730. [DOI] [PubMed] [Google Scholar]
- 26.Phillips RL, et al. , How other countries use deprivation indices—and why the United States desperately needs one. Health Affairs, 2016. 35(11): p. 1991–1998. [DOI] [PubMed] [Google Scholar]
- 27.Singh GK and Jemal A, Socioeconomic and racial/ethnic disparities in cancer mortality, incidence, and survival in the United States, 1950–2014: over six decades of changing patterns and widening inequalities. Journal of environmental and public health, 2017. 2017. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28.Kimbro RT, Sharp G, and Denney JT, Home and away: area socioeconomic disadvantage and obesity risk. Health & place, 2017. 44: p. 94–102. [DOI] [PubMed] [Google Scholar]
- 29.Census. American Community Survey Information Guide. 2017. [cited 2019 April 20]; Available from: https://www.census.gov/content/dam/Census/programs-surveys/acs/about/ACS_Information_Guide.pdf.
- 30.Centers for Disease Control and Prevention. 2000 CDC Growth Charts for the United States: Methods and Development. 2010. [cited 2015 17 July]; Available from: http://www.cdc.gov/nchs/data/series/sr_11/sr11_246.pdf.
- 31.Rogers W, Regression standard errors in clustered samples. Stata technical bulletin, 1994. 3(13). [Google Scholar]
- 32.Braveman P, Egerter S, and Williams DR, The social determinants of health: coming of age. Annual review of public health, 2011. 32: p. 381–398. [DOI] [PubMed] [Google Scholar]
- 33.Gottlieb LM, et al. , Moving electronic medical records upstream: incorporating social determinants of health. American Journal of Preventive Medicine, 2015. 48(2): p. 215–218. [DOI] [PubMed] [Google Scholar]
- 34.Li B, et al. , It Matters! Teaching Social Determinants of Health in the Intensive Care Unit to Healthcare Providers, in C40. CRITICAL CARE: THE ART OF WAR-INNOVATIONS IN EDUCATION. 2019, American Thoracic Society. p. A4784–A4784. [Google Scholar]
- 35.Robert SA, Socioeconomic position and health: the independent contribution of community socioeconomic context. Annual review of sociology, 1999. 25(1): p. 489–516. [Google Scholar]
- 36.Robert SA, Neighborhood socioeconomic context and adult health: The mediating role of individual health behaviors and psychosocial factors. Annals of the New York Academy of Sciences, 1999. 896(1): p. 465–468. [DOI] [PubMed] [Google Scholar]
- 37.Kawachi I and Berkman LF, Neighborhoods and health. 2003: Oxford University Press. [Google Scholar]
- 38.Macintyre S and Ellaway A, Neighborhoods and health: an overview. Neighborhoods and health, 2003: p. 20–42. [Google Scholar]
- 39.Adler NE and Rehkopf DH, US disparities in health: descriptions, causes, and mechanisms. Annu. Rev. Public Health, 2008. 29: p. 235–252. [DOI] [PubMed] [Google Scholar]
- 40.Perkins DD and Taylor RB, Ecological assessments of community disorder: Their relationship to fear of crime and theoretical implications. American journal of community psychology, 1996. 24(1): p. 63–107. [DOI] [PubMed] [Google Scholar]
- 41.Ross CE and Jang SJ, Neighborhood disorder, fear, and mistrust: The buffering role of social ties with neighbors. American journal of community psychology, 2000. 28(4): p. 401–420. [DOI] [PubMed] [Google Scholar]
- 42.Shavers VL, Measurement of socioeconomic status in health disparities research. Journal of the national medical association, 2007. 99(9): p. 1013. [PMC free article] [PubMed] [Google Scholar]
- 43.Cederberg M, Hartsmar N, and Lingärde S, Thematic report: Socioeconomic disadvantage. Report from the EPASI (Educational Policies that Address Social Inequality) project supported by the European Commission’s department of Education & Culture, SOCRATES programme, 2009. 2(2). [Google Scholar]
- 44.Sobal J and Stunkard AJ, Socioeconomic status and obesity: a review of the literature. Psychological bulletin, 1989. 105(2): p. 260. [DOI] [PubMed] [Google Scholar]
- 45.Parsons TJ, et al. , Childhood predictors of adult obesity: a systematic review. International journal of obesity, 1999. 23. [PubMed] [Google Scholar]
- 46.Glass TA and McAtee MJ, Behavioral science at the crossroads in public health: extending horizons, envisioning the future. Social science & medicine, 2006. 62(7): p. 1650–1671. [DOI] [PubMed] [Google Scholar]
- 47.Alvarado SE, Neighborhood disadvantage and obesity across childhood and adolescence: Evidence from the NLSY children and young adults cohort (1986–2010). Soc Sci Res, 2016. 57: p. 80–98. [DOI] [PubMed] [Google Scholar]
- 48.Twaits A and Alwan NA, The association between area-based deprivation and change in body-mass index over time in primary school children: a population-based cohort study in Hampshire, UK. Int J Obes (Lond), 2020. 44(3): p. 628–636. [DOI] [PubMed] [Google Scholar]
- 49.Rossen LM, Neighbourhood economic deprivation explains racial/ethnic disparities in overweight and obesity among children and adolescents in the U.S.A. J Epidemiol Community Health, 2014. 68(2): p. 123–9. [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Data Availability Statement
The datasets generated during and/or analyzed during the current study are available from the corresponding author on reasonable request.