In this issue of Blood, Williams et al,1 through a new report from the Childhood Cancer Survivor Study (CCSS), bring into clearer perspective the added burden of neurocognitive, psychosocial, and quality of life impairments to the well-recognized physiologic morbidity and premature mortality in long-term survivors of childhood Hodgkin lymphoma (HL). While there is some “bad news,” the authors did identify potentially modifiable factors that might prevent or mitigate adverse outcomes.
While childhood HL is a rare cancer in the United States (estimated at 1077 cases in those <20 years old in 2021), its high relative survival with treatment (current 5-year estimate is 94.3% for this age group) leads to a large population of survivors (2018 estimate is 40 851 survivors).2 Furthermore, HL is one of the top cancer types based on the proportion of survivors at ≥15 years after diagnosis.3 Thus, understanding and optimally managing the long-term impacts of treatment across a range of health4 as well as psychosocial and quality of life5 domains is of high clinical importance. These latter issues are comprehensively addressed in this new report from the CCSS, which focuses on HL survivors diagnosed from 1970 to 1999 along with a comparison group of siblings of the CCSS cohort. This research extends recent findings from the entire CCSS cohort that included many different types of childhood cancers.6-8 This report provides results specific to childhood HL and treatments used for HL (including those from the 1980s and 1990s).
First, some bad news. Compared with CCSS siblings, HL survivors reported greater neurocognitive impairments (most significantly for emotional regulation impairment, 11.5% vs 16.6%; and memory impairment, 5.7% vs 8.1%), and psychosocial impairments (most significantly for depression, 7.0% vs 9.1%). They reported lower quality of life across multiple domains (most significantly for general health, 9.7% vs 30.4%; and physical function, 3.0% vs 11.2%). HL survivors were also more likely to be unemployed (4% vs 10%).
Now for the good news. First, while the risk of neurocognitive, psychosocial, and quality of life impairments and lower social attainment was increased for the HL survivors relative to the CCSS siblings, the absolute risk must be kept in perspective in that the vast majority of HL survivors did not experience these impairments. Second, and of relevance to HL survivors who experience these impairments, this report found that the impairments were correlated with several modifiable risk factors. Specifically, never smoking, body mass index <25 kg/m2, and meeting Centers for Disease Control and Prevention guidelines for physical activity were associated with fewer impairments. Third, major HL treatments and treatment intensity were largely not associated with impairments in most of the domains assessed. Fourth, while several chronic health conditions, highlighted by hypertension and stroke, were associated with impairment in most of the neurocognitive, psychosocial, and quality of life domains, this does raise the potential that adequate management of these conditions could mitigate these outcomes. Finally, all associations between HL treatments and neurocognitive impairments were fully mediated by chronic health conditions.
There are many strengths of this report, including the use of HL participants from a defined survivor cohort; a large sample size of HL survivors and sibling controls; inclusion of participants diagnosed through 1999, which included HL treatments used up to this time point and of high relevance to many current long-term HL survivors; use of sibling controls for a comparison group; national scope of the study; use of comprehensive and validated measures; and sophisticated statistical analyses.
There are also limitations, 2 of which are worth noting here. First, the measure of neurocognitive impairment was self-reported and not objectively measured (eg, by neuropsychological testing). These 2 approaches often do not correlate well (eg, report of memory problems is not the same as actual poor recall on a memory test), and subjective measures are more likely to be an indicator of psychologic distress than cognitive impairment.9 Nevertheless, subjective neurocognitive impairments are still important to address clinically in patient management.
The other major limitation, and of greater consequence, is that this report uses a cross-sectional study design embedded in the CCSS cohort. As such, the associations reported here are cross-sectional correlations where cause and effect cannot be unambiguously determined, and many of the associations could go in the opposite direction (eg, memory impairment or poor quality of life lead to smoking, obesity, or lower physical activity). While there are other data supporting these modifiable factors improving neurocognitive, psychosocial, and quality of life in the general population, it will be important to test these associations in longitudinal studies and, most importantly, in intervention studies of HL survivors.
In summary, these data provide an important impetus for a well-designed trial in HL survivors, perhaps with an intervention that simultaneously incorporates the multiple modifiable factors identified here and that measures the positive impacts of changes in these lifestyle factors on neurocognitive and psychosocial functioning in the context of HL. But in the interim, there is sufficient evidence that smoking cessation, weight control, physical activity, and management of chronic diseases (eg, hypertension) all improve health and longevity in the general population and thus supports their thoughtful incorporation into HL survivorship plans.
Conflict-of-interest disclosure: J.R.C. has grant support from BMS, Genentech, and NanoString, all unrelated to this commentary.