Childhood Cancer Survivors, Late Effects, and a New Model for Understanding Survivorship

In 1950, at a time when cancer remission was measured in days, Farber stated that “The use of chemotherapeutic agents now available . . . should do much to reduce the number of instances of ‘incurable cancer’ in infants and children.”¹ Today, in 2007, approximately 80% of children with cancer are cured.²

However, it gradually was realized that this cure had a cost; ie, the curative therapy could damage a child's developing organ systems. Some problems, such as cognitive deficits following cranial radiotherapy, were apparent soon after completion of therapy. However, many sequelae were not recognized until survivors were a decade or more beyond their cancer. In 1974, Meadows and D’Angio described different methods and approaches to “detect the late effects of cancer therapy”³ ; thus began the expansion of the concept of cure to include long-term outcomes.

In this issue of JAMA, Geenen and colleagues⁴ provide a noteworthy and important contribution to understanding the long-term consequences of cancer therapy experienced by survivors of childhood cancer. Even though much research had been conducted in this area, resulting in many reports on the quality of life of long-term survivors or the prevalence of specific late effects, prior to 2006 there were no studies with adequate sample sizes that provided a composite estimate of the morbidity associated with therapy for childhood cancer. Last year, the report from the Childhood Cancer Survivor Study (CCSS) on chronic health conditions included more than 10 000 adult survivors of childhood cancer.⁵ These 2 studies complement each other, filling the gaps of respective study design limitations and providing a robust estimate of morbidity in this population of survivors, many of whom are now entering their young and mid-adult years.

The study by Geenen and colleagues,⁴ from the late-effects clinic (Polikliniek Late Effecten Kindertumorien) of the Emma Children's Hospital/Academic Medical Center (EKZ/AMC), Amsterdam, the Netherlands, has several strengths. For example, it is truly remarkable that among childhood cancer survivors treated from 1966-1996, only 1.5% were lost to follow-up. For perspective, in a report on long-term outcomes of survivors of acute lymphoblastic leukemia, Pui et al⁶ from St Jude Children's Research Hospital reported that 5.1% of survivors were lost to follow-up and another 16% did not respond to a short mailed questionnaire.⁶ One primary limitation in survivorship research is the potential bias of estimates introduced by such loss to follow-up or to nonparticipation. With 98.5% capture of data, this is not an issue for the EKZ/AMC study. Another major limitation in survivorship research is the frequent dependence on self-reported outcomes. Not only were the survivors in the EKZ/AMC study successfully contacted, but 94.3% were evaluated by a physician, including 79% of the cohort who were evaluated at a single institution using standardized protocols for screening and assessment. It is unlikely that any North American institution with a moderate to large population of cancer patients will be able to achieve this participation rate with clinical examinations.

The study by Geenen and colleagues also has several limitations. Most importantly, there is not a comparison group of healthy individuals to provide perspective or a background rate of the outcomes. Second, the population is a homogenous group of white Northern European cancer survivors with a relatively high standard of living and easy access to the health care system, thus affecting the generalizability of the findings to other populations. And third, the study reflects the outcomes of a cohort treated at a single, moderate-sized institution.

These strengths and limitations complement the study from the CCSS. Whereas the CCSS had a large and geographically, socioeconomically, and relatively racially diverse population diagnosed at 26 different institutions in North America and a comparison group of siblings, the outcomes were largely self-reported. In comparison with other large studies of survivors, the CCSS had a response rate of 81.2% for patients successfully contacted, but 14.6% of the eligible cohort was lost to follow-up.

Despite the differences in study design and populations, these 2 studies reported remarkably similar findings. For both studies, the median patient age at the time of study was approximately 25 years, and the interval from cancer diagnosis to study was about 17 years. Similarly, both studies reported that the majority of survivors had at least 1 chronic disease or adverse outcome (EKZ/AMC, 74.5%; CCSS, 62.3%) and that a significant proportion had a serious physical health problem (EKZ/AMC, 35.9%; CCSS, 27.5%). One potential explanation for the modestly higher prevalence of physical morbidity in the EKZ/AMC cohort in comparison with the CCSS cohort is that patient self-reported outcomes (as in the CCSS) likely underestimate the burden of morbidity because conditions that are often asymptomatic (eg, osteoporosis, insulin resistance, dyslipidemia, hypertension) may remain undiagnosed. In combination, the EKZ/AMC and CCSS studies provide confidence in the findings that childhood cancer survivors treated in the 1960s to the early 1990s are highly likely to develop at least 1 chronic disease, often severe in nature. Furthermore, many of these young adult survivors will have multiple problems involving several organ systems, thus hastening the aging process as this population enters their middle years.

This issue of JAMA is devoted to chronic diseases in infants, children, and young adults, and, as such, one of the most important diseases of children, cancer, is included. But childhood cancer, in and of itself, is rarely a chronic disease. Instead, the therapies used to cure the cancer can cause organ toxicity that manifests as a wide array of chronic diseases, becoming clinically evident at different periods of the survivor's life. In fact, cancer survivorship represents a different and unique model that is separate and distinct from the traditional models of chronic disease. In children with classic chronic diseases, such as diabetes, cystic fibrosis, and cerebral palsy, their primary illness extends into adulthood. The long-term outcomes of these chronic diseases are largely the result of a specific organ system malfunction early in life. A more analogous population is children with congenital heart disease. Although the primary problem is cured with an intervention, these children may experience health problems as adults secondary to their curative surgery. However, these problems are usually confined to the heart.

In contrast, children with cancer who are cured of their primary disease are often relatively asymptomatic during their adolescent years, and then develop one or more problems many years later distinct from their original disease. In some situations, this is a new disease that would not have occurred without the previous cancer therapy, such as a second sarcoma in the radiation field of an extremity⁷ or severe aortic stenosis following mediastinal radiotherapy.⁸ However, some of the diseases are likely an exaggeration of the survivor's genetics and lifestyle behaviors. For example, a young adult survivor of childhood leukemia who has a family history of obesity, insulin resistance, and premature cardiovascular disease and who has undergone cranial radiotherapy at a young age is much more likely to be obese⁹ and physically inactive¹⁰ than she or he would have been simply based on genetics and lifestyle behaviors. Thus, as an obese and sedentary young adult, she or he is more likely than siblings to develop insulin resistance at a younger age—and ultimately to have a myocardial infarction or cerebrovascular accident. Importantly, the risks and severity of many late effects, including cardiovascular disease, are modifiable.

Why is an understanding of this new cancer survivor paradigm important? At the time when most children and adolescents complete their therapy and are cured of their cancer, they generally do not have any apparent sequelae of their chemotherapy, surgery, or radiation. In many countries, including the United States, Canada, and the United Kingdom, most survivors are gradually lost to follow-up and are not managed by clinicians at their treating institutions.^{11 - 12} Most adult survivors of childhood cancer do not have a summary of their cancer and their cancer therapy, do not know what therapy they received other than in broad terms, and are not aware of their risks for sequelae.¹³ Realizing that survivors of pediatric cancer comprise a relatively small percentage of the population (1 of every 640 young adults in the United States is a pediatric cancer survivor¹⁴ ), most primary care physicians, as well as surgeons, obstetricians, cardiologists, and other specialists, are not familiar with the health risks of this relatively heterogeneous population. Hence, as these survivors enter their young and mid-adult years, a period when their risk for many serious late effects is at its highest, they might not be screened for various late effects that may be modifiable or amenable to early diagnosis and treatment. Moreover, preventive interventions, such as avoidance of tobacco, increasing levels of physical activity, or promoting a healthy diet, often are not being delivered. In addition, survivors are developing a wide array of problems and presenting with new signs and symptoms to primary care physicians and specialists who generally are unfamiliar with the relationship of different treatment exposures to different diseases.

The traditional chronic disease model is inadequate to describe the evolution of morbidity and mortality in survivors of childhood cancer. Based on an evolving understanding of the cancer survivor model, it is important to promote the concept of risk-based health care of survivors. Such an approach should be longitudinal, proactive, and anticipatory and should include a systematic plan of prevention and surveillance based on risks associated with the previous cancer, cancer therapy, genetic predispositions, lifestyle behaviors, and comorbid health conditions.¹⁵ To assist physicians and other health care professionals in this process, a few groups, including the 240-institution Children's Oncology Group, have developed recommendations for screening and surveillance of childhood cancer survivors,^{16 - 18} as well as guidelines for follow-up care of childhood cancer.¹⁹

In summary, Geenen and colleagues provide provocative data highlighting the excessive and severe morbidity that follows therapy for childhood cancer, generally becoming evident in the young and mid-adult years. What the future holds for survivors who have bravely won the battle with their childhood cancer is uncertain. It is critically important for physicians to recognize these risks, facilitate risk-based health care, and strive to improve therapy that not only improves cure rates but also reduces long-term morbidity. Furthermore, it is essential to conduct research to prospectively monitor, document, and quantify risks as the population of childhood cancer survivors ages.