New Research Pinpoints Potential Risk Factors for Late Deaths in Childhood Cancer Survivors


A cohort study identified factors associated with late deaths in childhood cancer survivors.

Childhood cancer survivors who grew up in a socioeconomically disadvantaged neighborhood, or who live with modifiable chronic health conditions, are at an elevated risk of death 5 years following their diagnosis, according to findings from a cohort study which were published in JAMA Oncology.

The presence of 1 or more modifiable chronic health condition (CHC) of grade 2 or higher and having frailty were associated with significant increases in all-cause and health-related late mortality, defined as death occurring at least 5 years after diagnosis. Survivors living in the most disadvantaged Census blocks were at a 5-fold to 8-fold increased risk for all-cause death compared with those living in the least disadvantaged Census block.

Investigators determined that survivors were at 7.6-fold significantly increased risk for all-cause mortality (standardized mortality ratio [SMR], 7.6; 95% CI, 7.2-8.1) and health-related late mortality (SMR, 7.6; 95% CI, 7.0-8.2). Survivors who had at least 10 CHCs of grade 1 to 4, or at least 3 or more CHCs of grade 3 to 4 at baseline had the highest risk. Modifiable CHCs were defined as modifiable conditions such as dyslipidemia, hypertension, diabetes, underweight or obesity, bone mineral deficiency, hypogonadism, hypothyroidism, and adrenal insufficiency that are associated with increased mortality in the general population and for which treatment and interventions were readily available and could conceivably alter the risk of late death.

Survivors of childhood cancer are known to have individual-level disadvantages such as lower educational attainment, unemployment or employment in lower-skilled jobs, inadequate insurance coverage, and lower income compared with their peers. Investigators initiated this longitudinal cohort study to assess the association between potentially modifiable CHCs and cancer mortality within the context of social determinants of health.

Investigators included all 9440 eligible survivors regardless of their participation in the St Jude Lifetime Cohort (SJLIFE) retrospective cohort study or their vital status to assess late mortality without bias. They then restricted evaluation of the associations between modifiable health conditions, social determinants, and late mortality, to 3407 adult study participants evaluated at St Jude and for whom data on modifiable health conditions and social determinants were available.

SJLIFE includes patients diagnosed and treated at St Jude Children’s Research Hospital from 1962 to 2012 who survived 5 or more years after diagnosis.

For the research cohort, investigators geocoded last-reported residential addresses to determine neighborhood-level socioeconomic status (SES) by US Census blocks using the area deprivation index (ADI), a composite measure derived from components of the American Community Survey reflective of 17 neighborhood-level SES measures including household income, employment status, and educational level. Each block was previously assigned a national percentile, ranking minimum SES disadvantage in the 1st percentile and maximum disadvantage in the 100th percentile. For 4.9% of on-campus participants, the ADI was designated as unassigned due to the unavailability of geocoding information such as post office boxes or international addresses.

Participants in the research cohort were assigned a healthy lifestyle index score using a composite of smoking status (ever or current vs never), alcohol consumption (heavy or risky drinking [>4 drinks/day or >14 drinks/week for men and >3 drinks/day or >7 drinks/week for women] vs no heavy or risky drinking), physical activity (meeting Centers for Disease Control and Prevention [CDC] guidelines [≥75 minutes of vigorous activity or ≥150 minutes of moderate and/or vigorous combined activity] vs not meeting CDC guidelines), and body mass index (calculated as weight in kilograms divided by height in meters squared; 18.5 to <25.0 vs other body mass index categories). Each domain of the healthy lifestyle index was scored 1 if healthy and 0 if otherwise, with scores of 4 assigned to those considered healthy in all 4 domains and 0 assigned to those considered unhealthy in all 4 domains. Frailty was defined using the Fried criteria.

Participants with 2 of the following conditions were defined as having prefrailty: low muscle mass, weakness, slow walking speed, self-reported exhaustion, or low energy expenditure. Participants with 3 or more of these conditions were defined as having frailty.

For the entire cohort, the median age at assessment was 27.5 years (range, 5.3-71.9) and the median duration of follow-up was 18.8 years (range, 5.0-58.0). Most survivors were male (55.2%) and of non-Hispanic White race and ethnicity (75.3%). Among adult on-campus participants, the median age at assessment was 35.4 years (range, 17.9-69.8) and the median duration of follow-up was 27.3 years (range, 7.3-54.7). Most participants were male (52.5%) and of non-Hispanic White race and ethnicity (81.7%).

Five to 9 years after diagnosis, the highest rate of death was due to recurrent cancer (9.2 deaths/1000 person-years; 95% CI, 8.3-10.1). In contrast, health-related causes of death exceeded deaths due to recurrence beginning at 10 to 14 years after diagnosis and beyond (2.8 deaths/1000 person-years at 10-14 years after diagnosis to 19.4 deaths/1000 person-years at ≥40 years after diagnosis). The rate of subsequent neoplasm–related deaths ranged from 1.9 per 1000 person-years (5-9 years after diagnosis) to 2.3 per 1000 person-years (20-24 years after diagnosis) between 5 and 29 years after diagnosis but increased to 8.8 per 1000 person-years by 40 or more years after diagnosis. The cumulative incidence of health-related causes of death exceeded cancer recurrence or progression by 25 years after diagnosis.

Among specific health-related causes of death, SMRs were 16.0 (95% CI, 14.4-17.8) for subsequent neoplasms, 4.2 (95% CI, 3.3-5.3) for cardiac causes, 9.0 (95% CI, 6.5-12.0) for pulmonary causes, and 4.3 (95% CI, 3.7-5.0) for other health-related causes.

Modifiable Risk Factors for Death

After adjusting for age during follow-up, age at diagnosis, sex, race and ethnicity, treatment exposures, annual household income, insurance status, and healthy lifestyle index score, investigators found that having 1 modifiable CHC of grade 2 or higher (rate ratios [RR], 2.2; 95% CI, 1.2-4.0; P = .01), 2 modifiable CHCs of grade 2 or higher (RR, 2.6; 95% CI, 1.4-4.9; P = .003), or 3 modifiable CHCs of grade 2 or higher (RR, 3.6; 95% CI, 1.8-7.1; P < .001) was associated with increased risk for late all-cause death.

Living in a US Census block with an ADI in the 51st to 80th percentile (RR, 5.5; 95% CI, 1.3-23.5; P = .02), an ADI in the 81st to 100th percentile (RR, 8.7; 95% CI, 2.0-37.6; P = .004), or an unassigned ADI (RR, 15.7; 95% CI, 3.5-70.3; P < .001) were associated with significant increases in the risk for late all-cause death. Similarly, frailty (RR, 2.3; 95% CI, 1.3-3.9; P = .004) was associated with risk for late all-cause death.

Cumulative mortality 8 years after baseline assessment was 13.2% (95% CI, 10.2%-16.3%) in survivors with 10 or more CHCs of grade 1 to 4 at baseline compared with 1.6% (95% CI, 0.6%-2.6%) in those with 0 to 5 CHCs of grade 1 to 4 at baseline and 2.6% (95% CI, 1.1%-4.1%) in those with 6 to 9 CHCs of grade 1 to 4 at baseline. Furthermore, investigators determined that the proportion of health-related causes of death increased with the number of CHCs identified at baseline.


Ehrhardt MJ, Liu Q, Dixon SB, et al. Association of modifiable health conditions and social determinants of health with late mortality in survivors of childhood cancer. JAMA Netw Open. 2023;6(2):e2255395. doi:10.1001/jamanetworkopen.2022.55395

Recent Videos
More Than A “Side Effect”: Why Pediatric Hearing Loss Is So Detrimental
How Can Writing Help Process Cancer-Related Grief?
Related Content
© 2024 MJH Life Sciences

All rights reserved.